MyArxiv
Computation and Language 33
☆ Auditing Gender Presentation Differences in Text-to-Image Models
Text-to-image models, which can generate high-quality images based on textual input, have recently enabled various content-creation tools. Despite significantly affecting a wide range of downstream applications, the distributions of these generated images are still not fully understood, especially when it comes to the potential stereotypical attributes of different genders. In this work, we propose a paradigm (Gender Presentation Differences) that utilizes fine-grained self-presentation attributes to study how gender is presented differently in text-to-image models. By probing gender indicators in the input text (e.g., "a woman" or "a man"), we quantify the frequency differences of presentation-centric attributes (e.g., "a shirt" and "a dress") through human annotation and introduce a novel metric: GEP. Furthermore, we propose an automatic method to estimate such differences. The automatic GEP metric based on our approach yields a higher correlation with human annotations than that based on existing CLIP scores, consistently across three state-of-the-art text-to-image models. Finally, we demonstrate the generalization ability of our metrics in the context of gender stereotypes related to occupations.
comment: Preprint, 23 pages, 14 figures
☆ Hard Prompts Made Easy: Gradient-Based Discrete Optimization for Prompt Tuning and Discovery
The strength of modern generative models lies in their ability to be controlled through text-based prompts. Typical "hard" prompts are made from interpretable words and tokens, and must be hand-crafted by humans. There are also "soft" prompts, which consist of continuous feature vectors. These can be discovered using powerful optimization methods, but they cannot be easily interpreted, re-used across models, or plugged into a text-based interface. We describe an approach to robustly optimize hard text prompts through efficient gradient-based optimization. Our approach automatically generates hard text-based prompts for both text-to-image and text-to-text applications. In the text-to-image setting, the method creates hard prompts for diffusion models, allowing API users to easily generate, discover, and mix and match image concepts without prior knowledge on how to prompt the model. In the text-to-text setting, we show that hard prompts can be automatically discovered that are effective in tuning LMs for classification.
comment: 14 pages, 10 figures, Code is available at \url{https://github.com/YuxinWenRick/hard-prompts-made-easy}
☆ Exploitation and exploration in text evolution. Quantifying planning and translation flows during writing
Writing is a complex process at the center of much of modern human activity. Despite it appears to be a linear process, writing conceals many highly non-linear processes. Previous research has focused on three phases of writing: planning, translation and transcription, and revision. While research has shown these are non-linear, they are often treated linearly when measured. Here, we introduce measures to detect and quantify subcycles of planning (exploration) and translation (exploitation) during the writing process. We apply these to a novel dataset that recorded the creation of a text in all its phases, from early attempts to the finishing touches on a final version. This dataset comes from a series of writing workshops in which, through innovative versioning software, we were able to record all the steps in the construction of a text. More than 60 junior researchers in science wrote a scientific essay intended for a general readership. We recorded each essay as a writing cloud, defined as a complex topological structure capturing the history of the essay itself. Through this unique dataset of writing clouds, we expose a representation of the writing process that quantifies its complexity and the writer's efforts throughout the draft and through time. Interestingly, this representation highlights the phases of "translation flow", where authors improve existing ideas, and exploration, where creative deviations appear as the writer returns to the planning phase. These turning points between translation and exploration become rarer as the writing process progresses and the author approaches the final version. Our results and the new measures introduced have the potential to foster the discussion about the non-linear nature of writing and support the development of tools that can support more creative and impactful writing processes.
☆ CALaMo: a Constructionist Assessment of Language Models
This paper presents a novel framework for evaluating Neural Language Models' linguistic abilities using a constructionist approach. Not only is the usage-based model in line with the underlying stochastic philosophy of neural architectures, but it also allows the linguist to keep meaning as a determinant factor in the analysis. We outline the framework and present two possible scenarios for its application.
☆ Efficiently Upgrading Multilingual Machine Translation Models to Support More Languages EACL 2023
With multilingual machine translation (MMT) models continuing to grow in size and number of supported languages, it is natural to reuse and upgrade existing models to save computation as data becomes available in more languages. However, adding new languages requires updating the vocabulary, which complicates the reuse of embeddings. The question of how to reuse existing models while also making architectural changes to provide capacity for both old and new languages has also not been closely studied. In this work, we introduce three techniques that help speed up effective learning of the new languages and alleviate catastrophic forgetting despite vocabulary and architecture mismatches. Our results show that by (1) carefully initializing the network, (2) applying learning rate scaling, and (3) performing data up-sampling, it is possible to exceed the performance of a same-sized baseline model with 30% computation and recover the performance of a larger model trained from scratch with over 50% reduction in computation. Furthermore, our analysis reveals that the introduced techniques help learn the new directions more effectively and alleviate catastrophic forgetting at the same time. We hope our work will guide research into more efficient approaches to growing languages for these MMT models and ultimately maximize the reuse of existing models.
comment: Accepted to EACL 2023 (Main)
☆ A Survey on Arabic Named Entity Recognition: Past, Recent Advances, and Future Trends
As more and more Arabic texts emerged on the Internet, extracting important information from these Arabic texts is especially useful. As a fundamental technology, Named entity recognition (NER) serves as the core component in information extraction technology, while also playing a critical role in many other Natural Language Processing (NLP) systems, such as question answering and knowledge graph building. In this paper, we provide a comprehensive review of the development of Arabic NER, especially the recent advances in deep learning and pre-trained language model. Specifically, we first introduce the background of Arabic NER, including the characteristics of Arabic and existing resources for Arabic NER. Then, we systematically review the development of Arabic NER methods. Traditional Arabic NER systems focus on feature engineering and designing domain-specific rules. In recent years, deep learning methods achieve significant progress by representing texts via continuous vector representations. With the growth of pre-trained language model, Arabic NER yields better performance. Finally, we conclude the method gap between Arabic NER and NER methods from other languages, which helps outline future directions for Arabic NER.
comment: arXiv admin note: text overlap with arXiv:2210.09263 by other authors
☆ Cluster-Level Contrastive Learning for Emotion Recognition in Conversations
A key challenge for Emotion Recognition in Conversations (ERC) is to distinguish semantically similar emotions. Some works utilise Supervised Contrastive Learning (SCL) which uses categorical emotion labels as supervision signals and contrasts in high-dimensional semantic space. However, categorical labels fail to provide quantitative information between emotions. ERC is also not equally dependent on all embedded features in the semantic space, which makes the high-dimensional SCL inefficient. To address these issues, we propose a novel low-dimensional Supervised Cluster-level Contrastive Learning (SCCL) method, which first reduces the high-dimensional SCL space to a three-dimensional affect representation space Valence-Arousal-Dominance (VAD), then performs cluster-level contrastive learning to incorporate measurable emotion prototypes. To help modelling the dialogue and enriching the context, we leverage the pre-trained knowledge adapters to infuse linguistic and factual knowledge. Experiments show that our method achieves new state-of-the-art results with 69.81% on IEMOCAP, 65.7% on MELD, and 62.51% on DailyDialog datasets. The analysis also proves that the VAD space is not only suitable for ERC but also interpretable, with VAD prototypes enhancing its performance and stabilising the training of SCCL. In addition, the pre-trained knowledge adapters benefit the performance of the utterance encoder and SCCL. Our code is available at: https://github.com/SteveKGYang/SCCL
comment: Accepted by IEEE Transactions on Affective Computing
☆ Learning Translation Quality Evaluation on Low Resource Languages from Large Language Models
Learned metrics such as BLEURT have in recent years become widely employed to evaluate the quality of machine translation systems. Training such metrics requires data which can be expensive and difficult to acquire, particularly for lower-resource languages. We show how knowledge can be distilled from Large Language Models (LLMs) to improve upon such learned metrics without requiring human annotators, by creating synthetic datasets which can be mixed into existing datasets, requiring only a corpus of text in the target language. We show that the performance of a BLEURT-like model on lower resource languages can be improved in this way.
☆ Natural Language Processing for Policymaking
Language is the medium for many political activities, from campaigns to news reports. Natural language processing (NLP) uses computational tools to parse text into key information that is needed for policymaking. In this chapter, we introduce common methods of NLP, including text classification, topic modeling, event extraction, and text scaling. We then overview how these methods can be used for policymaking through four major applications including data collection for evidence-based policymaking, interpretation of political decisions, policy communication, and investigation of policy effects. Finally, we highlight some potential limitations and ethical concerns when using NLP for policymaking. This text is from Chapter 7 (pages 141-162) of the Handbook of Computational Social Science for Policy (2023). Open Access on Springer: https://doi.org/10.1007/978-3-031-16624-2
comment: Handbook of Computational Social Science for Policy (2023), Chapter 7 (pages 141-162). Open Access on Springer: https://doi.org/10.1007/978-3-031-16624-2
☆ Entity-Aware Dual Co-Attention Network for Fake News Detection EACL 2023
Fake news and misinformation spread rapidly on the Internet. How to identify it and how to interpret the identification results have become important issues. In this paper, we propose a Dual Co-Attention Network (Dual-CAN) for fake news detection, which takes news content, social media replies, and external knowledge into consideration. Our experimental results support that the proposed Dual-CAN outperforms current representative models in two benchmark datasets. We further make in-depth discussions by comparing how models work in both datasets with empirical analysis of attention weights.
comment: EACL 2023 Findings
☆ What do Language Models know about word senses? Zero-Shot WSD with Language Models and Domain Inventories
Language Models are the core for almost any Natural Language Processing system nowadays. One of their particularities is their contextualized representations, a game changer feature when a disambiguation between word senses is necessary. In this paper we aim to explore to what extent language models are capable of discerning among senses at inference time. We performed this analysis by prompting commonly used Languages Models such as BERT or RoBERTa to perform the task of Word Sense Disambiguation (WSD). We leverage the relation between word senses and domains, and cast WSD as a textual entailment problem, where the different hypothesis refer to the domains of the word senses. Our results show that this approach is indeed effective, close to supervised systems.
comment: Presented at GWC2023
☆ The Effect of Metadata on Scientific Literature Tagging: A Cross-Field Cross-Model Study WWW 2023
Due to the exponential growth of scientific publications on the Web, there is a pressing need to tag each paper with fine-grained topics so that researchers can track their interested fields of study rather than drowning in the whole literature. Scientific literature tagging is beyond a pure multi-label text classification task because papers on the Web are prevalently accompanied by metadata information such as venues, authors, and references, which may serve as additional signals to infer relevant tags. Although there have been studies making use of metadata in academic paper classification, their focus is often restricted to one or two scientific fields (e.g., computer science and biomedicine) and to one specific model. In this work, we systematically study the effect of metadata on scientific literature tagging across 19 fields. We select three representative multi-label classifiers (i.e., a bag-of-words model, a sequence-based model, and a pre-trained language model) and explore their performance change in scientific literature tagging when metadata are fed to the classifiers as additional features. We observe some ubiquitous patterns of metadata's effects across all fields (e.g., venues are consistently beneficial to paper tagging in almost all cases), as well as some unique patterns in fields other than computer science and biomedicine, which are not explored in previous studies.
comment: 11 pages; Accepted to WWW 2023
☆ Learning Manner of Execution from Partial Corrections
Some actions must be executed in different ways depending on the context. For example, wiping away marker requires vigorous force while wiping away almonds requires more gentle force. In this paper we provide a model where an agent learns which manner of action execution to use in which context, drawing on evidence from trial and error and verbal corrections when it makes a mistake (e.g., ``no, gently''). The learner starts out with a domain model that lacks the concepts denoted by the words in the teacher's feedback; both the words describing the context (e.g., marker) and the adverbs like ``gently''. We show that through the the semantics of coherence, our agent can perform the symbol grounding that's necessary for exploiting the teacher's feedback so as to solve its domain-level planning problem: to perform its actions in the current context in the right way.
☆ AutoWS: Automated Weak Supervision Framework for Text Classification
Creating large, good quality labeled data has become one of the major bottlenecks for developing machine learning applications. Multiple techniques have been developed to either decrease the dependence of labeled data (zero/few-shot learning, weak supervision) or to improve the efficiency of labeling process (active learning). Among those, Weak Supervision has been shown to reduce labeling costs by employing hand crafted labeling functions designed by domain experts. We propose AutoWS -- a novel framework for increasing the efficiency of weak supervision process while decreasing the dependency on domain experts. Our method requires a small set of labeled examples per label class and automatically creates a set of labeling functions to assign noisy labels to numerous unlabeled data. Noisy labels can then be aggregated into probabilistic labels used by a downstream discriminative classifier. Our framework is fully automatic and requires no hyper-parameter specification by users. We compare our approach with different state-of-the-art work on weak supervision and noisy training. Experimental results show that our method outperforms competitive baselines.
☆ PLACES: Prompting Language Models for Social Conversation Synthesis EACL 2023
Collecting high quality conversational data can be very expensive for most applications and infeasible for others due to privacy, ethical, or similar concerns. A promising direction to tackle this problem is to generate synthetic dialogues by prompting large language models. In this work, we use a small set of expert-written conversations as in-context examples to synthesize a social conversation dataset using prompting. We perform several thorough evaluations of our synthetic conversations compared to human-collected conversations. This includes various dimensions of conversation quality with human evaluation directly on the synthesized conversations, and interactive human evaluation of chatbots fine-tuned on the synthetically generated dataset. We additionally demonstrate that this prompting approach is generalizable to multi-party conversations, providing potential to create new synthetic data for multi-party tasks. Our synthetic multi-party conversations were rated more favorably across all measured dimensions compared to conversation excerpts sampled from a human-collected multi-party dataset.
comment: In EACL 2023. 25 pages, 4 figures, 26 tables. Link to code forthcoming
☆ Continual Learning of Language Models ICLR 2023
Language models (LMs) have been instrumental for the rapid advance of natural language processing. This paper studies continual learning of LMs, in particular, continual domain-adaptive pre-training (or continual DAP-training). Existing research has shown that further pre-training an LM using a domain corpus to adapt the LM to the domain can improve the end-task performance in the domain. This paper proposes a novel method to continually DAP-train an LM with a sequence of unlabeled domain corpora to adapt the LM to these domains to improve their end-task performances. The key novelty of our method is a soft-masking mechanism that directly controls the update to the LM. A novel proxy is also proposed to preserve the general knowledge in the original LM. Additionally, it contrasts the representations of the previously learned domain knowledge (including the general knowledge in the pre-trained LM) and the knowledge from the current full network to achieve knowledge integration. The method not only overcomes catastrophic forgetting, but also achieves knowledge transfer to improve end-task performances. Empirical evaluation demonstrates the effectiveness of the proposed method.
comment: ICLR 2023
☆ Bringing the State-of-the-Art to Customers: A Neural Agent Assistant Framework for Customer Service Support EMNLP 2022
Building Agent Assistants that can help improve customer service support requires inputs from industry users and their customers, as well as knowledge about state-of-the-art Natural Language Processing (NLP) technology. We combine expertise from academia and industry to bridge the gap and build task/domain-specific Neural Agent Assistants (NAA) with three high-level components for: (1) Intent Identification, (2) Context Retrieval, and (3) Response Generation. In this paper, we outline the pipeline of the NAA's core system and also present three case studies in which three industry partners successfully adapt the framework to find solutions to their unique challenges. Our findings suggest that a collaborative process is instrumental in spurring the development of emerging NLP models for Conversational AI tasks in industry. The full reference implementation code and results are available at \url{https://github.com/VectorInstitute/NAA}
comment: Camera Ready Version of Paper Published in EMNLP 2022 Industry Track
☆ An entity-guided text summarization framework with relational heterogeneous graph neural network
Two crucial issues for text summarization to generate faithful summaries are to make use of knowledge beyond text and to make use of cross-sentence relations in text. Intuitive ways for the two issues are Knowledge Graph (KG) and Graph Neural Network (GNN) respectively. Entities are semantic units in text and in KG. This paper focuses on both issues by leveraging entities mentioned in text to connect GNN and KG for summarization. Firstly, entities are leveraged to construct a sentence-entity graph with weighted multi-type edges to model sentence relations, and a relational heterogeneous GNN for summarization is proposed to calculate node encodings. Secondly, entities are leveraged to link the graph to KG to collect knowledge. Thirdly, entities guide a two-step summarization framework defining a multi-task selector to select salient sentences and entities, and using an entity-focused abstractor to compress the sentences. GNN is connected with KG by constructing sentence-entity graphs where entity-entity edges are built based on KG, initializing entity embeddings on KG, and training entity embeddings using entity-entity edges. The relational heterogeneous GNN utilizes both edge weights and edge types in GNN to calculate graphs with weighted multi-type edges. Experiments show the proposed method outperforms extractive baselines including the HGNN-based HGNNSum and abstractive baselines including the entity-driven SENECA on CNN/DM, and outperforms most baselines on NYT50. Experiments on sub-datasets show the density of sentence-entity edges greatly influences the performance of the proposed method. The greater the density, the better the performance. Ablations show effectiveness of the method.
comment: 7 tables, 5 figures
☆ Exploring the Benefits of Training Expert Language Models over Instruction Tuning
Recently, Language Models (LMs) instruction-tuned on multiple tasks, also known as multitask-prompted fine-tuning (MT), have shown the capability to generalize to unseen tasks. Previous work has shown that scaling the number of training tasks is the key component in making stronger MT LMs. In this work, we report an unexpected finding that an expert LM fine-tuned on just a single task can outperform an MT LM trained with 300+ different tasks on 11 different unseen datasets and on 13 datasets of the BIG-bench benchmark by a mean accuracy of 3.20% and 1.29%, respectively. This finding casts doubt on the previously held belief that simply scaling the number of tasks makes stronger MT LMs. Leveraging this finding, we further show that this distributed approach of training a separate expert LM per training task instead of a single MT LM for zero-shot inference possesses many benefits including (1) avoiding negative task transfer that often occurs during instruction tuning, (2) being able to continually learn new tasks without having to re-train on previous tasks to avoid catastrophic forgetting, and (3) showing compositional capabilities when merging individual experts together. The code is available at https://github.com/joeljang/ELM.
☆ UDApter -- Efficient Domain Adaptation Using Adapters
We propose two methods to make unsupervised domain adaptation (UDA) more parameter efficient using adapters, small bottleneck layers interspersed with every layer of the large-scale pre-trained language model (PLM). The first method deconstructs UDA into a two-step process: first by adding a domain adapter to learn domain-invariant information and then by adding a task adapter that uses domain-invariant information to learn task representations in the source domain. The second method jointly learns a supervised classifier while reducing the divergence measure. Compared to strong baselines, our simple methods perform well in natural language inference (MNLI) and the cross-domain sentiment classification task. We even outperform unsupervised domain adaptation methods such as DANN and DSN in sentiment classification, and we are within 0.85% F1 for natural language inference task, by fine-tuning only a fraction of the full model parameters. We release our code at https://github.com/declare-lab/UDAPTER
☆ Capturing Topic Framing via Masked Language Modeling EMNLP 2022
Differential framing of issues can lead to divergent world views on important issues. This is especially true in domains where the information presented can reach a large audience, such as traditional and social media. Scalable and reliable measurement of such differential framing is an important first step in addressing them. In this work, based on the intuition that framing affects the tone and word choices in written language, we propose a framework for modeling the differential framing of issues through masked token prediction via large-scale fine-tuned language models (LMs). Specifically, we explore three key factors for our framework: 1) prompt generation methods for the masked token prediction; 2) methods for normalizing the output of fine-tuned LMs; 3) robustness to the choice of pre-trained LMs used for fine-tuning. Through experiments on a dataset of articles from traditional media outlets covering five diverse and politically polarized topics, we show that our framework can capture differential framing of these topics with high reliability.
comment: In Findings of EMNLP 2022
♻ ☆ A Non-monotonic Self-terminating Language Model ICLR 2023
Recent large-scale neural autoregressive sequence models have shown impressive performances on a variety of natural language generation tasks. However, their generated sequences often exhibit degenerate properties such as non-termination, undesirable repetition, and premature termination, when generated with decoding algorithms such as greedy search, beam search, top-$k$ sampling, and nucleus sampling. In this paper, we focus on the problem of non-terminating sequences resulting from an incomplete decoding algorithm. We first define an incomplete probable decoding algorithm which includes greedy search, top-$k$ sampling, and nucleus sampling, beyond the incomplete decoding algorithm originally put forward by Welleck et al. (2020). We then propose a non-monotonic self-terminating language model, which significantly relaxes the constraint of monotonically increasing termination probability in the originally proposed self-terminating language model by Welleck et al. (2020), to address the issue of non-terminating sequences when using incomplete probable decoding algorithms. We prove that our proposed model prevents non-terminating sequences when using not only incomplete probable decoding algorithms but also beam search. We empirically validate our model on sequence completion tasks with various architectures.
comment: Published as a conference paper at ICLR 2023
♻ ☆ N-Gram Nearest Neighbor Machine Translation
Nearest neighbor machine translation augments the Autoregressive Translation~(AT) with $k$-nearest-neighbor retrieval, by comparing the similarity between the token-level context representations of the target tokens in the query and the datastore. However, the token-level representation may introduce noise when translating ambiguous words, or fail to provide accurate retrieval results when the representation generated by the model contains indistinguishable context information, e.g., Non-Autoregressive Translation~(NAT) models. In this paper, we propose a novel $n$-gram nearest neighbor retrieval method that is model agnostic and applicable to both AT and NAT models. Specifically, we concatenate the adjacent $n$-gram hidden representations as the key, while the tuple of corresponding target tokens is the value. In inference, we propose tailored decoding algorithms for AT and NAT models respectively. We demonstrate that the proposed method consistently outperforms the token-level method on both AT and NAT models as well on general as on domain adaptation translation tasks. On domain adaptation, the proposed method brings $1.03$ and $2.76$ improvements regarding the average BLEU score on AT and NAT models respectively.
♻ ☆ FADO: Feedback-Aware Double COntrolling Network for Emotional Support Conversation SC
Emotional Support Conversation (ESConv) aims to reduce help-seekers'emotional distress with the supportive strategy and response. It is essential for the supporter to select an appropriate strategy with the feedback of the help-seeker (e.g., emotion change during dialog turns, etc) in ESConv. However, previous methods mainly focus on the dialog history to select the strategy and ignore the help-seeker's feedback, leading to the wrong and user-irrelevant strategy prediction. In addition, these approaches only model the context-to-strategy flow and pay less attention to the strategy-to-context flow that can focus on the strategy-related context for generating the strategy-constrain response. In this paper, we propose a Feedback-Aware Double COntrolling Network (FADO) to make a strategy schedule and generate the supportive response. The core module in FADO consists of a dual-level feedback strategy selector and a double control reader. Specifically, the dual-level feedback strategy selector leverages the turn-level and conversation-level feedback to encourage or penalize strategies. The double control reader constructs the novel strategy-to-context flow for generating the strategy-constrain response. Furthermore, a strategy dictionary is designed to enrich the semantic information of the strategy and improve the quality of strategy-constrain response. Experimental results on ESConv show that the proposed FADO has achieved the state-of-the-art performance in terms of both strategy selection and response generation. Our code is available at https://github.com/Thedatababbler/FADO.
comment: Accepted on Knowl. Based Syst. (SCI I)
♻ ☆ Recent Advances in Neural Text Generation: A Task-Agnostic Survey
In recent years much effort has been devoted to applying neural models to the task of natural language generation. The challenge is to generate natural human-like text, and to control the generation process. This paper presents a task-agnostic survey of recent advances in neural text generation. These advances have been achieved by numerous developments, which we group under the following four headings: data construction, neural frameworks, training and inference strategies, and evaluation metrics. Finally we discuss the future directions for the development of neural text generation including neural pipelines and exploiting back-ground knowledge.
♻ ☆ Scaling Back-Translation with Domain Text Generation for Sign Language Gloss Translation EACL 2023
Sign language gloss translation aims to translate the sign glosses into spoken language texts, which is challenging due to the scarcity of labeled gloss-text parallel data. Back translation (BT), which generates pseudo-parallel data by translating in-domain spoken language texts into sign glosses, has been applied to alleviate the data scarcity problem. However, the lack of large-scale high-quality domain spoken language text data limits the effect of BT. In this paper, to overcome the limitation, we propose a Prompt based domain text Generation (PGEN) approach to produce the large-scale in-domain spoken language text data. Specifically, PGEN randomly concatenates sentences from the original in-domain spoken language text data as prompts to induce a pre-trained language model (i.e., GPT-2) to generate spoken language texts in a similar style. Experimental results on three benchmarks of sign language gloss translation in varied languages demonstrate that BT with spoken language texts generated by PGEN significantly outperforms the compared methods. In addition, as the scale of spoken language texts generated by PGEN increases, the BT technique can achieve further improvements, demonstrating the effectiveness of our approach. We release the code and data for facilitating future research in this field.
comment: Accepted at EACL 2023 (main conference)
♻ ☆ Less is More: Understanding Word-level Textual Adversarial Attack via n-gram Frequency Descend
Word-level textual adversarial attacks have achieved striking performance in fooling natural language processing models. However, the fundamental questions of why these attacks are effective, and the intrinsic properties of the adversarial examples (AEs), are still not well understood. This work attempts to interpret textual attacks through the lens of $n$-gram frequency. Specifically, it is revealed that existing word-level attacks exhibit a strong tendency toward generation of examples with $n$-gram frequency descend ($n$-FD). Intuitively, this finding suggests a natural way to improve model robustness by training the model on the $n$-FD examples. To verify this idea, we devise a model-agnostic and gradient-free AE generation approach that relies solely on the $n$-gram frequency information, and further integrate it into the recently proposed convex hull framework for adversarial training. Surprisingly, the resultant method performs quite similarly to the original gradient-based method in terms of model robustness. These findings provide a human-understandable perspective for interpreting word-level textual adversarial attacks, and a new direction to improve model robustness.
comment: 8 pages, 4 figures. In progress
♻ ☆ Prompting Neural Machine Translation with Translation Memories AAAI 2023
Improving machine translation (MT) systems with translation memories (TMs) is of great interest to practitioners in the MT community. However, previous approaches require either a significant update of the model architecture and/or additional training efforts to make the models well-behaved when TMs are taken as additional input. In this paper, we present a simple but effective method to introduce TMs into neural machine translation (NMT) systems. Specifically, we treat TMs as prompts to the NMT model at test time, but leave the training process unchanged. The result is a slight update of an existing NMT system, which can be implemented in a few hours by anyone who is familiar with NMT. Experimental results on several datasets demonstrate that our system significantly outperforms strong baselines.
comment: Accepted to AAAI 2023
♻ ☆ A Comprehensive Comparison of Pre-training Language Models
Recently, the development of pre-trained language models has brought natural language processing (NLP) tasks to the new state-of-the-art. In this paper we explore the efficiency of various pre-trained language models. We pre-train a list of transformer-based models with the same amount of text and the same training steps. The experimental results shows that the most improvement upon the origin BERT is adding the RNN-layer to capture more contextual information for short text understanding. But the conclusion is: There are no remarkable improvement for short text understanding for similar BERT structures. Data-centric method[12] can achieve better performance.
♻ ☆ Don't Blame the Annotator: Bias Already Starts in the Annotation Instructions EACL 2023
In recent years, progress in NLU has been driven by benchmarks. These benchmarks are typically collected by crowdsourcing, where annotators write examples based on annotation instructions crafted by dataset creators. In this work, we hypothesize that annotators pick up on patterns in the crowdsourcing instructions, which bias them to write many similar examples that are then over-represented in the collected data. We study this form of bias, termed instruction bias, in 14 recent NLU benchmarks, showing that instruction examples often exhibit concrete patterns, which are propagated by crowdworkers to the collected data. This extends previous work (Geva et al., 2019) and raises a new concern of whether we are modeling the dataset creator's instructions, rather than the task. Through a series of experiments, we show that, indeed, instruction bias can lead to overestimation of model performance, and that models struggle to generalize beyond biases originating in the crowdsourcing instructions. We further analyze the influence of instruction bias in terms of pattern frequency and model size, and derive concrete recommendations for creating future NLU benchmarks.
comment: Accepted to EACL 2023
♻ ☆ MetaQA: Combining Expert Agents for Multi-Skill Question Answering EACL 2023
The recent explosion of question answering (QA) datasets and models has increased the interest in the generalization of models across multiple domains and formats by either training on multiple datasets or by combining multiple models. Despite the promising results of multi-dataset models, some domains or QA formats may require specific architectures, and thus the adaptability of these models might be limited. In addition, current approaches for combining models disregard cues such as question-answer compatibility. In this work, we propose to combine expert agents with a novel, flexible, and training-efficient architecture that considers questions, answer predictions, and answer-prediction confidence scores to select the best answer among a list of answer candidates. Through quantitative and qualitative experiments we show that our model i) creates a collaboration between agents that outperforms previous multi-agent and multi-dataset approaches in both in-domain and out-of-domain scenarios, ii) is highly data-efficient to train, and iii) can be adapted to any QA format. We release our code and a dataset of answer predictions from expert agents for 16 QA datasets to foster future developments of multi-agent systems on https://github.com/UKPLab/MetaQA.
comment: Accepted at EACL 2023
♻ ☆ MCoNaLa: A Benchmark for Code Generation from Multiple Natural Languages
While there has been a recent burgeoning of applications at the intersection of natural and programming languages, such as code generation and code summarization, these applications are usually English-centric. This creates a barrier for program developers who are not proficient in English. To mitigate this gap in technology development across languages, we propose a multilingual dataset, MCoNaLa, to benchmark code generation from natural language commands extending beyond English. Modeled off of the methodology from the English Code/Natural Language Challenge (CoNaLa) dataset, we annotated a total of 896 NL-code pairs in three languages: Spanish, Japanese, and Russian. We present a quantitative evaluation of performance on the MCoNaLa dataset by testing with state-of-the-art code generation systems. While the difficulties vary across these three languages, all systems lag significantly behind their English counterparts, revealing the challenges in adapting code generation to new languages.
♻ ☆ ProKD: An Unsupervised Prototypical Knowledge Distillation Network for Zero-Resource Cross-Lingual Named Entity Recognition AAAI 2023
For named entity recognition (NER) in zero-resource languages, utilizing knowledge distillation methods to transfer language-independent knowledge from the rich-resource source languages to zero-resource languages is an effective means. Typically, these approaches adopt a teacher-student architecture, where the teacher network is trained in the source language, and the student network seeks to learn knowledge from the teacher network and is expected to perform well in the target language. Despite the impressive performance achieved by these methods, we argue that they have two limitations. Firstly, the teacher network fails to effectively learn language-independent knowledge shared across languages due to the differences in the feature distribution between the source and target languages. Secondly, the student network acquires all of its knowledge from the teacher network and ignores the learning of target language-specific knowledge. Undesirably, these limitations would hinder the model's performance in the target language. This paper proposes an unsupervised prototype knowledge distillation network (ProKD) to address these issues. Specifically, ProKD presents a contrastive learning-based prototype alignment method to achieve class feature alignment by adjusting the distance among prototypes in the source and target languages, boosting the teacher network's capacity to acquire language-independent knowledge. In addition, ProKD introduces a prototypical self-training method to learn the intrinsic structure of the language by retraining the student network on the target data using samples' distance information from prototypes, thereby enhancing the student network's ability to acquire language-specific knowledge. Extensive experiments on three benchmark cross-lingual NER datasets demonstrate the effectiveness of our approach.
comment: AAAI 2023
Computer Vision and Pattern Recognition 76
☆ How Reliable is Your Regression Model's Uncertainty Under Real-World Distribution Shifts?
Many important computer vision applications are naturally formulated as regression problems. Within medical imaging, accurate regression models have the potential to automate various tasks, helping to lower costs and improve patient outcomes. Such safety-critical deployment does however require reliable estimation of model uncertainty, also under the wide variety of distribution shifts that might be encountered in practice. Motivated by this, we set out to investigate the reliability of regression uncertainty estimation methods under various real-world distribution shifts. To that end, we propose an extensive benchmark of 8 image-based regression datasets with different types of challenging distribution shifts. We then employ our benchmark to evaluate many of the most common uncertainty estimation methods, as well as two state-of-the-art uncertainty scores from the task of out-of-distribution detection. We find that while methods are well calibrated when there is no distribution shift, they all become highly overconfident on many of the benchmark datasets. This uncovers important limitations of current uncertainty estimation methods, and the proposed benchmark therefore serves as a challenge to the research community. We hope that our benchmark will spur more work on how to develop truly reliable regression uncertainty estimation methods. Code is available at https://github.com/fregu856/regression_uncertainty.
comment: Code is available at https://github.com/fregu856/regression_uncertainty
☆ Auditing Gender Presentation Differences in Text-to-Image Models
Text-to-image models, which can generate high-quality images based on textual input, have recently enabled various content-creation tools. Despite significantly affecting a wide range of downstream applications, the distributions of these generated images are still not fully understood, especially when it comes to the potential stereotypical attributes of different genders. In this work, we propose a paradigm (Gender Presentation Differences) that utilizes fine-grained self-presentation attributes to study how gender is presented differently in text-to-image models. By probing gender indicators in the input text (e.g., "a woman" or "a man"), we quantify the frequency differences of presentation-centric attributes (e.g., "a shirt" and "a dress") through human annotation and introduce a novel metric: GEP. Furthermore, we propose an automatic method to estimate such differences. The automatic GEP metric based on our approach yields a higher correlation with human annotations than that based on existing CLIP scores, consistently across three state-of-the-art text-to-image models. Finally, we demonstrate the generalization ability of our metrics in the context of gender stereotypes related to occupations.
comment: Preprint, 23 pages, 14 figures
☆ HumanMAC: Masked Motion Completion for Human Motion Prediction
Human motion prediction is a classical problem in computer vision and computer graphics, which has a wide range of practical applications. Previous effects achieve great empirical performance based on an encoding-decoding fashion. The methods of this fashion work by first encoding previous motions to latent representations and then decoding the latent representations into predicted motions. However, in practice, they are still unsatisfactory due to several issues, including complicated loss constraints, cumbersome training processes, and scarce switch of different categories of motions in prediction. In this paper, to address the above issues, we jump out of the foregoing fashion and propose a novel framework from a new perspective. Specifically, our framework works in a denoising diffusion style. In the training stage, we learn a motion diffusion model that generates motions from random noise. In the inference stage, with a denoising procedure, we make motion prediction conditioning on observed motions to output more continuous and controllable predictions. The proposed framework enjoys promising algorithmic properties, which only needs one loss in optimization and is trained in an end-to-end manner. Additionally, it accomplishes the switch of different categories of motions effectively, which is significant in realistic tasks, \textit{e.g.}, the animation task. Comprehensive experiments on benchmarks confirm the superiority of the proposed framework. The project page is available at \url{https://lhchen.top/Human-MAC}.
☆ Toward Face Biometric De-identification using Adversarial Examples AAAI-23
The remarkable success of face recognition (FR) has endangered the privacy of internet users particularly in social media. Recently, researchers turned to use adversarial examples as a countermeasure. In this paper, we assess the effectiveness of using two widely known adversarial methods (BIM and ILLC) for de-identifying personal images. We discovered, unlike previous claims in the literature, that it is not easy to get a high protection success rate (suppressing identification rate) with imperceptible adversarial perturbation to the human visual system. Finally, we found out that the transferability of adversarial examples is highly affected by the training parameters of the network with which they are generated.
comment: Accepted at the AAAI-23 workshop on Artificial Intelligence for Cyber Security (AICS)
☆ Deep Class-Incremental Learning: A Survey
Deep models, e.g., CNNs and Vision Transformers, have achieved impressive achievements in many vision tasks in the closed world. However, novel classes emerge from time to time in our ever-changing world, requiring a learning system to acquire new knowledge continually. For example, a robot needs to understand new instructions, and an opinion monitoring system should analyze emerging topics every day. Class-Incremental Learning (CIL) enables the learner to incorporate the knowledge of new classes incrementally and build a universal classifier among all seen classes. Correspondingly, when directly training the model with new class instances, a fatal problem occurs -- the model tends to catastrophically forget the characteristics of former ones, and its performance drastically degrades. There have been numerous efforts to tackle catastrophic forgetting in the machine learning community. In this paper, we survey comprehensively recent advances in deep class-incremental learning and summarize these methods from three aspects, i.e., data-centric, model-centric, and algorithm-centric. We also provide a rigorous and unified evaluation of 16 methods in benchmark image classification tasks to find out the characteristics of different algorithms empirically. Furthermore, we notice that the current comparison protocol ignores the influence of memory budget in model storage, which may result in unfair comparison and biased results. Hence, we advocate fair comparison by aligning the memory budget in evaluation, as well as several memory-agnostic performance measures. The source code to reproduce these evaluations is available at https://github.com/zhoudw-zdw/CIL_Survey/
comment: Code is available at https://github.com/zhoudw-zdw/CIL_Survey/
☆ S4R: Self-Supervised Semantic Scene Reconstruction from RGB-D Scans
Most deep learning approaches to comprehensive semantic modeling of 3D indoor spaces require costly dense annotations in the 3D domain. In this work, we explore a central 3D scene modeling task, namely, semantic scene reconstruction, using a fully self-supervised approach. To this end, we design a trainable model that employs both incomplete 3D reconstructions and their corresponding source RGB-D images, fusing cross-domain features into volumetric embeddings to predict complete 3D geometry, color, and semantics. Our key technical innovation is to leverage differentiable rendering of color and semantics, using the observed RGB images and a generic semantic segmentation model as color and semantics supervision, respectively. We additionally develop a method to synthesize an augmented set of virtual training views complementing the original real captures, enabling more efficient self-supervision for semantics. In this work we propose an end-to-end trainable solution jointly addressing geometry completion, colorization, and semantic mapping from a few RGB-D images, without 3D or 2D ground-truth. Our method is the first, to our knowledge, fully self-supervised method addressing completion and semantic segmentation of real-world 3D scans. It performs comparably well with the 3D supervised baselines, surpasses baselines with 2D supervision on real datasets, and generalizes well to unseen scenes.
☆ Ethical Considerations for Collecting Human-Centric Image Datasets
Human-centric image datasets are critical to the development of computer vision technologies. However, recent investigations have foregrounded significant ethical issues related to privacy and bias, which have resulted in the complete retraction, or modification, of several prominent datasets. Recent works have tried to reverse this trend, for example, by proposing analytical frameworks for ethically evaluating datasets, the standardization of dataset documentation and curation practices, privacy preservation methodologies, as well as tools for surfacing and mitigating representational biases. Little attention, however, has been paid to the realities of operationalizing ethical data collection. To fill this gap, we present a set of key ethical considerations and practical recommendations for collecting more ethically-minded human-centric image data. Our research directly addresses issues of privacy and bias by contributing to the research community best practices for ethical data collection, covering purpose, privacy and consent, as well as diversity. We motivate each consideration by drawing on lessons from current practices, dataset withdrawals and audits, and analytical ethical frameworks. Our research is intended to augment recent scholarship, representing an important step toward more responsible data curation practices.
☆ Pole Estimation and Optical Navigation using Circle of Latitude Projections
Images of both rotating celestial bodies (e.g., asteroids) and spheroidal planets with banded atmospheres (e.g., Jupiter) can contain features that are well-modeled as a circle of latitude (CoL). The projections of these CoLs appear as ellipses in images collected by cameras or telescopes onboard exploration spacecraft. This work shows how CoL projections may be used to determine the pole orientation and covariance for a spinning asteroid. In the case of a known planet modeled as an oblate spheroid, it is shown how similar CoL projections may be used for spacecraft localization. These methods are developed using the principles of projective geometry. Numerical results are provided for simulated images of asteroid Bennu (for pole orientation) and of Jupiter (for spacecraft localization).
☆ NICER-SLAM: Neural Implicit Scene Encoding for RGB SLAM
Neural implicit representations have recently become popular in simultaneous localization and mapping (SLAM), especially in dense visual SLAM. However, previous works in this direction either rely on RGB-D sensors, or require a separate monocular SLAM approach for camera tracking and do not produce high-fidelity dense 3D scene reconstruction. In this paper, we present NICER-SLAM, a dense RGB SLAM system that simultaneously optimizes for camera poses and a hierarchical neural implicit map representation, which also allows for high-quality novel view synthesis. To facilitate the optimization process for mapping, we integrate additional supervision signals including easy-to-obtain monocular geometric cues and optical flow, and also introduce a simple warping loss to further enforce geometry consistency. Moreover, to further boost performance in complicated indoor scenes, we also propose a local adaptive transformation from signed distance functions (SDFs) to density in the volume rendering equation. On both synthetic and real-world datasets we demonstrate strong performance in dense mapping, tracking, and novel view synthesis, even competitive with recent RGB-D SLAM systems.
comment: Video: https://youtu.be/tUXzqEZWg2w
☆ Local Neural Descriptor Fields: Locally Conditioned Object Representations for Manipulation ICRA 2023
A robot operating in a household environment will see a wide range of unique and unfamiliar objects. While a system could train on many of these, it is infeasible to predict all the objects a robot will see. In this paper, we present a method to generalize object manipulation skills acquired from a limited number of demonstrations, to novel objects from unseen shape categories. Our approach, Local Neural Descriptor Fields (L-NDF), utilizes neural descriptors defined on the local geometry of the object to effectively transfer manipulation demonstrations to novel objects at test time. In doing so, we leverage the local geometry shared between objects to produce a more general manipulation framework. We illustrate the efficacy of our approach in manipulating novel objects in novel poses -- both in simulation and in the real world.
comment: ICRA 2023, Project Page: https://elchun.github.io/lndf/
☆ A Deep Learning-based in silico Framework for Optimization on Retinal Prosthetic Stimulation
We propose a neural network-based framework to optimize the perceptions simulated by the in silico retinal implant model pulse2percept. The overall pipeline consists of a trainable encoder, a pre-trained retinal implant model and a pre-trained evaluator. The encoder is a U-Net, which takes the original image and outputs the stimulus. The pre-trained retinal implant model is also a U-Net, which is trained to mimic the biomimetic perceptual model implemented in pulse2percept. The evaluator is a shallow VGG classifier, which is trained with original images. Based on 10,000 test images from the MNIST dataset, we show that the convolutional neural network-based encoder performs significantly better than the trivial downsampling approach, yielding a boost in the weighted F1-Score by 36.17% in the pre-trained classifier with 6x10 electrodes. With this fully neural network-based encoder, the quality of the downstream perceptions can be fine-tuned using gradient descent in an end-to-end fashion.
☆ Look around and learn: self-improving object detection by exploration
Object detectors often experience a drop in performance when new environmental conditions are insufficiently represented in the training data. This paper studies how to automatically fine-tune a pre-existing object detector while exploring and acquiring images in a new environment without relying on human intervention, i.e., in an utterly self-supervised fashion. In our setting, an agent initially learns to explore the environment using a pre-trained off-the-shelf detector to locate objects and associate pseudo-labels. By assuming that pseudo-labels for the same object must be consistent across different views, we learn an exploration policy mining hard samples and we devise a novel mechanism for producing refined predictions from the consensus among observations. Our approach outperforms the current state-of-the-art, and it closes the performance gap against a fully supervised setting without relying on ground-truth annotations. We also compare various exploration policies for the agent to gather more informative observations. Code and dataset will be made available upon paper acceptance
☆ PhysFormer++: Facial Video-based Physiological Measurement with SlowFast Temporal Difference Transformer
Remote photoplethysmography (rPPG), which aims at measuring heart activities and physiological signals from facial video without any contact, has great potential in many applications (e.g., remote healthcare and affective computing). Recent deep learning approaches focus on mining subtle rPPG clues using convolutional neural networks with limited spatio-temporal receptive fields, which neglect the long-range spatio-temporal perception and interaction for rPPG modeling. In this paper, we propose two end-to-end video transformer based architectures, namely PhysFormer and PhysFormer++, to adaptively aggregate both local and global spatio-temporal features for rPPG representation enhancement. As key modules in PhysFormer, the temporal difference transformers first enhance the quasi-periodic rPPG features with temporal difference guided global attention, and then refine the local spatio-temporal representation against interference. To better exploit the temporal contextual and periodic rPPG clues, we also extend the PhysFormer to the two-pathway SlowFast based PhysFormer++ with temporal difference periodic and cross-attention transformers. Furthermore, we propose the label distribution learning and a curriculum learning inspired dynamic constraint in frequency domain, which provide elaborate supervisions for PhysFormer and PhysFormer++ and alleviate overfitting. Comprehensive experiments are performed on four benchmark datasets to show our superior performance on both intra- and cross-dataset testings. Unlike most transformer networks needed pretraining from large-scale datasets, the proposed PhysFormer family can be easily trained from scratch on rPPG datasets, which makes it promising as a novel transformer baseline for the rPPG community.
comment: Accepted by International Journal of Computer Vision (IJCV). arXiv admin note: substantial text overlap with arXiv:2111.12082
☆ Aligning Multi-Sequence CMR Towards Fully Automated Myocardial Pathology Segmentation
Myocardial pathology segmentation (MyoPS) is critical for the risk stratification and treatment planning of myocardial infarction (MI). Multi-sequence cardiac magnetic resonance (MS-CMR) images can provide valuable information. For instance, balanced steady-state free precession cine sequences present clear anatomical boundaries, while late gadolinium enhancement and T2-weighted CMR sequences visualize myocardial scar and edema of MI, respectively. Existing methods usually fuse anatomical and pathological information from different CMR sequences for MyoPS, but assume that these images have been spatially aligned. However, MS-CMR images are usually unaligned due to the respiratory motions in clinical practices, which poses additional challenges for MyoPS. This work presents an automatic MyoPS framework for unaligned MS-CMR images. Specifically, we design a combined computing model for simultaneous image registration and information fusion, which aggregates multi-sequence features into a common space to extract anatomical structures (i.e., myocardium). Consequently, we can highlight the informative regions in the common space via the extracted myocardium to improve MyoPS performance, considering the spatial relationship between myocardial pathologies and myocardium. Experiments on a private MS-CMR dataset and a public dataset from the MYOPS2020 challenge show that our framework could achieve promising performance for fully automatic MyoPS.
☆ Revisiting Pre-training in Audio-Visual Learning
Pre-training technique has gained tremendous success in enhancing model performance on various tasks, but found to perform worse than training from scratch in some uni-modal situations. This inspires us to think: are the pre-trained models always effective in the more complex multi-modal scenario, especially for the heterogeneous modalities such as audio and visual ones? We find that the answer is No. Specifically, we explore the effects of pre-trained models on two audio-visual learning scenarios: cross-modal initialization and multi-modal joint learning. When cross-modal initialization is applied, the phenomena of "dead channel" caused by abnormal Batchnorm parameters hinders the utilization of model capacity. Thus, we propose Adaptive Batchnorm Re-initialization (ABRi) to better exploit the capacity of pre-trained models for target tasks. In multi-modal joint learning, we find a strong pre-trained uni-modal encoder would bring negative effects on the encoder of another modality. To alleviate such problem, we introduce a two-stage Fusion Tuning strategy, taking better advantage of the pre-trained knowledge while making the uni-modal encoders cooperate with an adaptive masking method. The experiment results show that our methods could further exploit pre-trained models' potential and boost performance in audio-visual learning.
☆ Structured Generative Models for Scene Understanding
This position paper argues for the use of \emph{structured generative models} (SGMs) for scene understanding. This requires the reconstruction of a 3D scene from an input image, whereby the contents of the image are causally explained in terms of models of instantiated objects, each with their own type, shape, appearance and pose, along with global variables like scene lighting and camera parameters. This approach also requires scene models which account for the co-occurrences and inter-relationships of objects in a scene. The SGM approach has the merits that it is compositional and generative, which lead to interpretability. To pursue the SGM agenda, we need models for objects and scenes, and approaches to carry out inference. We first review models for objects, which include ``things'' (object categories that have a well defined shape), and ``stuff'' (categories which have amorphous spatial extent). We then move on to review \emph{scene models} which describe the inter-relationships of objects. Perhaps the most challenging problem for SGMs is \emph{inference} of the objects, lighting and camera parameters, and scene inter-relationships from input consisting of a single or multiple images. We conclude with a discussion of issues that need addressing to advance the SGM agenda.
comment: 33 pages, 10 figures
☆ Explainable Action Prediction through Self-Supervision on Scene Graphs ICRA
This work explores scene graphs as a distilled representation of high-level information for autonomous driving, applied to future driver-action prediction. Given the scarcity and strong imbalance of data samples, we propose a self-supervision pipeline to infer representative and well-separated embeddings. Key aspects are interpretability and explainability; as such, we embed in our architecture attention mechanisms that can create spatial and temporal heatmaps on the scene graphs. We evaluate our system on the ROAD dataset against a fully-supervised approach, showing the superiority of our training regime.
comment: Accepted to the 2023 IEEE International Conference on Robotics and Automation (ICRA)
☆ VertXNet: An Ensemble Method for Vertebrae Segmentation and Identification of Spinal X-Ray
Reliable vertebrae annotations are key to perform analysis of spinal X-ray images. However, obtaining annotation of vertebrae from those images is usually carried out manually due to its complexity (i.e. small structures with varying shape), making it a costly and tedious process. To accelerate this process, we proposed an ensemble pipeline, VertXNet, that combines two state-of-the-art (SOTA) segmentation models (respectively U-Net and Mask R-CNN) to automatically segment and label vertebrae in X-ray spinal images. Moreover, VertXNet introduces a rule-based approach that allows to robustly infer vertebrae labels (by locating the 'reference' vertebrae which are easier to segment than others) for a given spinal X-ray image. We evaluated the proposed pipeline on three spinal X-ray datasets (two internal and one publicly available), and compared against vertebrae annotated by radiologists. Our experimental results have shown that the proposed pipeline outperformed two SOTA segmentation models on our test dataset (MEASURE 1) with a mean Dice of 0.90, vs. a mean Dice of 0.73 for Mask R-CNN and 0.72 for U-Net. To further evaluate the generalization ability of VertXNet, the pre-trained pipeline was directly tested on two additional datasets (PREVENT and NHANES II) and consistent performance was observed with a mean Dice of 0.89 and 0.88, respectively. Overall, VertXNet demonstrated significantly improved performance for vertebra segmentation and labeling for spinal X-ray imaging, and evaluation on both in-house clinical trial data and publicly available data further proved its generalization.
☆ Med-NCA: Robust and Lightweight Segmentation with Neural Cellular Automata
Access to the proper infrastructure is critical when performing medical image segmentation with Deep Learning. This requirement makes it difficult to run state-of-the-art segmentation models in resource-constrained scenarios like primary care facilities in rural areas and during crises. The recently emerging field of Neural Cellular Automata (NCA) has shown that locally interacting one-cell models can achieve competitive results in tasks such as image generation or segmentations in low-resolution inputs. However, they are constrained by high VRAM requirements and the difficulty of reaching convergence for high-resolution images. To counteract these limitations we propose Med-NCA, an end-to-end NCA training pipeline for high-resolution image segmentation. Our method follows a two-step process. Global knowledge is first communicated between cells across the downscaled image. Following that, patch-based segmentation is performed. Our proposed Med-NCA outperforms the classic UNet by 2% and 3% Dice for hippocampus and prostate segmentation, respectively, while also being 500 times smaller. We also show that Med-NCA is by design invariant with respect to image scale, shape and translation, experiencing only slight performance degradation even with strong shifts; and is robust against MRI acquisition artefacts. Med-NCA enables high-resolution medical image segmentation even on a Raspberry Pi B+, arguably the smallest device able to run PyTorch and that can be powered by a standard power bank.
☆ OSRT: Omnidirectional Image Super-Resolution with Distortion-aware Transformer
Omnidirectional images (ODIs) have obtained lots of research interest for immersive experiences. Although ODIs require extremely high resolution to capture details of the entire scene, the resolutions of most ODIs are insufficient. Previous methods attempt to solve this issue by image super-resolution (SR) on equirectangular projection (ERP) images. However, they omit geometric properties of ERP in the degradation process, and their models can hardly generalize to real ERP images. In this paper, we propose Fisheye downsampling, which mimics the real-world imaging process and synthesizes more realistic low-resolution samples. Then we design a distortion-aware Transformer (OSRT) to modulate ERP distortions continuously and self-adaptively. Without a cumbersome process, OSRT outperforms previous methods by about 0.2dB on PSNR. Moreover, we propose a convenient data augmentation strategy, which synthesizes pseudo ERP images from plain images. This simple strategy can alleviate the over-fitting problem of large networks and significantly boost the performance of ODISR. Extensive experiments have demonstrated the state-of-the-art performance of our OSRT. Codes and models will be available at https://github.com/Fanghua-Yu/OSRT.
comment: main paper + supplement
☆ Using t-distributed stochastic neighbor embedding for visualization and segmentation of 3D point clouds of plants
In this work, the use of t-SNE is proposed to embed 3D point clouds of plants into 2D space for plant characterization. It is demonstrated that t-SNE operates as a practical tool to flatten and visualize a complete 3D plant model in 2D space. The perplexity parameter of t-SNE allows 2D rendering of plant structures at various organizational levels. Aside from the promise of serving as a visualization tool for plant scientists, t-SNE also provides a gateway for processing 3D point clouds of plants using their embedded counterparts in 2D. In this paper, simple methods were proposed to perform semantic segmentation and instance segmentation via grouping the embedded 2D points. The evaluation of these methods on a public 3D plant data set conveys the potential of t-SNE for enabling of 2D implementation of various steps involved in automatic 3D phenotyping pipelines.
☆ SimCon Loss with Multiple Views for Text Supervised Semantic Segmentation
Learning to segment images purely by relying on the image-text alignment from web data can lead to sub-optimal performance due to noise in the data. The noise comes from the samples where the associated text does not correlate with the image's visual content. Instead of purely relying on the alignment from the noisy data, this paper proposes a novel loss function termed SimCon, which accounts for intra-modal similarities to determine the appropriate set of positive samples to align. Further, using multiple views of the image (created synthetically) for training and combining the SimCon loss with it makes the training more robust. This version of the loss is termed MV-SimCon. The empirical results demonstrate that using the proposed loss function leads to consistent improvements on zero-shot, text supervised semantic segmentation and outperforms state-of-the-art by $+3.0\%$, $+3.3\%$ and $+6.9\%$ on PASCAL VOC, PASCAL Context and MSCOCO, respectively. With test time augmentations, we set a new record by improving these results further to $58.7\%$, $26.6\%$, and $33.3\%$ on PASCAL VOC, PASCAL Context, and MSCOCO, respectively. In addition, using the proposed loss function leads to robust training and faster convergence.
☆ High-Resolution GAN Inversion for Degraded Images in Large Diverse Datasets AAAI2023
The last decades are marked by massive and diverse image data, which shows increasingly high resolution and quality. However, some images we obtained may be corrupted, affecting the perception and the application of downstream tasks. A generic method for generating a high-quality image from the degraded one is in demand. In this paper, we present a novel GAN inversion framework that utilizes the powerful generative ability of StyleGAN-XL for this problem. To ease the inversion challenge with StyleGAN-XL, Clustering \& Regularize Inversion (CRI) is proposed. Specifically, the latent space is firstly divided into finer-grained sub-spaces by clustering. Instead of initializing the inversion with the average latent vector, we approximate a centroid latent vector from the clusters, which generates an image close to the input image. Then, an offset with a regularization term is introduced to keep the inverted latent vector within a certain range. We validate our CRI scheme on multiple restoration tasks (i.e., inpainting, colorization, and super-resolution) of complex natural images, and show preferable quantitative and qualitative results. We further demonstrate our technique is robust in terms of data and different GAN models. To our best knowledge, we are the first to adopt StyleGAN-XL for generating high-quality natural images from diverse degraded inputs. Code is available at https://github.com/Booooooooooo/CRI.
comment: Accepted by AAAI2023
☆ AniPixel: Towards Animatable Pixel-Aligned Human Avatar
Neural radiance field using pixel-aligned features can render photo-realistic novel views. However, when pixel-aligned features are directly introduced to human avatar reconstruction, the rendering can only be conducted for still humans, rather than animatable avatars. In this paper, we propose AniPixel, a novel animatable and generalizable human avatar reconstruction method that leverages pixel-aligned features for body geometry prediction and RGB color blending. Technically, to align the canonical space with the target space and the observation space, we propose a bidirectional neural skinning field based on skeleton-driven deformation to establish the target-to-canonical and canonical-to-observation correspondences. Then, we disentangle the canonical body geometry into a normalized neutral-sized body and a subject-specific residual for better generalizability. As the geometry and appearance are closely related, we introduce pixel-aligned features to facilitate the body geometry prediction and detailed surface normals to reinforce the RGB color blending. Moreover, we devise a pose-dependent and view direction-related shading module to represent the local illumination variance. Experiments show that our AniPixel renders comparable novel views while delivering better novel pose animation results than state-of-the-art methods. The code will be released.
☆ PAMI: partition input and aggregate outputs for model interpretation
There is an increasing demand for interpretation of model predictions especially in high-risk applications. Various visualization approaches have been proposed to estimate the part of input which is relevant to a specific model prediction. However, most approaches require model structure and parameter details in order to obtain the visualization results, and in general much effort is required to adapt each approach to multiple types of tasks particularly when model backbone and input format change over tasks. In this study, a simple yet effective visualization framework called PAMI is proposed based on the observation that deep learning models often aggregate features from local regions for model predictions. The basic idea is to mask majority of the input and use the corresponding model output as the relative contribution of the preserved input part to the original model prediction. For each input, since only a set of model outputs are collected and aggregated, PAMI does not require any model detail and can be applied to various prediction tasks with different model backbones and input formats. Extensive experiments on multiple tasks confirm the proposed method performs better than existing visualization approaches in more precisely finding class-specific input regions, and when applied to different model backbones and input formats. The source code will be released publicly.
comment: 28 figures
☆ 3D Vessel Segmentation with Limited Guidance of 2D Structure-agnostic Vessel Annotations
Delineating 3D blood vessels is essential for clinical diagnosis and treatment, however, is challenging due to complex structure variations and varied imaging conditions. Supervised deep learning has demonstrated its superior capacity in automatic 3D vessel segmentation. However, the reliance on expensive 3D manual annotations and limited capacity for annotation reuse hinder the clinical applications of supervised models. To avoid the repetitive and laborious annotating and make full use of existing vascular annotations, this paper proposes a novel 3D shape-guided local discrimination model for 3D vascular segmentation under limited guidance from public 2D vessel annotations. The primary hypothesis is that 3D vessels are composed of semantically similar voxels and exhibit tree-shaped morphology. Accordingly, the 3D region discrimination loss is firstly proposed to learn the discriminative representation measuring voxel-wise similarities and cluster semantically consistent voxels to form the candidate 3D vascular segmentation in unlabeled images; secondly, based on the similarity of the tree-shaped morphology between 2D and 3D vessels, the Crop-and-Overlap strategy is presented to generate reference masks from 2D structure-agnostic vessel annotations, which are fit for varied vascular structures, and the adversarial loss is introduced to guide the tree-shaped morphology of 3D vessels; thirdly, the temporal consistency loss is proposed to foster the training stability and keep the model updated smoothly. To further enhance the model's robustness and reliability, the orientation-invariant CNN module and Reliability-Refinement algorithm are presented. Experimental results from the public 3D cerebrovascular and 3D arterial tree datasets demonstrate that our model achieves comparable effectiveness against nine supervised models.
comment: Submitted to IEEE TMI Journal
☆ Boosting Zero-shot Classification with Synthetic Data Diversity via Stable Diffusion
Recent research has shown it is possible to perform zero-shot classification tasks by training a classifier with synthetic data generated by a diffusion model. However, the performance of this approach is still inferior to that of recent vision-language models. It has been suggested that the reason for this is a domain gap between the synthetic and real data. In our work, we show that this domain gap is not the main issue, and that diversity in the synthetic dataset is more important. We propose a \textit{bag of tricks} to improve diversity and are able to achieve performance on par with one of the vision-language models, CLIP. More importantly, this insight allows us to endow zero-shot classification capabilities on any classification model.
comment: (7 pages, 3 figures, 2 tables, preprint)
☆ Multi-organ segmentation: a progressive exploration of learning paradigms under scarce annotation
Precise delineation of multiple organs or abnormal regions in the human body from medical images plays an essential role in computer-aided diagnosis, surgical simulation, image-guided interventions, and especially in radiotherapy treatment planning. Thus, it is of great significance to explore automatic segmentation approaches, among which deep learning-based approaches have evolved rapidly and witnessed remarkable progress in multi-organ segmentation. However, obtaining an appropriately sized and fine-grained annotated dataset of multiple organs is extremely hard and expensive. Such scarce annotation limits the development of high-performance multi-organ segmentation models but promotes many annotation-efficient learning paradigms. Among these, studies on transfer learning leveraging external datasets, semi-supervised learning using unannotated datasets and partially-supervised learning integrating partially-labeled datasets have led the dominant way to break such dilemma in multi-organ segmentation. We first review the traditional fully supervised method, then present a comprehensive and systematic elaboration of the 3 abovementioned learning paradigms in the context of multi-organ segmentation from both technical and methodological perspectives, and finally summarize their challenges and future trends.
comment: 23 pages, 4 figures, 5 tables
☆ Fine-grained Affordance Annotation for Egocentric Hand-Object Interaction Videos WACV
Object affordance is an important concept in hand-object interaction, providing information on action possibilities based on human motor capacity and objects' physical property thus benefiting tasks such as action anticipation and robot imitation learning. However, the definition of affordance in existing datasets often: 1) mix up affordance with object functionality; 2) confuse affordance with goal-related action; and 3) ignore human motor capacity. This paper proposes an efficient annotation scheme to address these issues by combining goal-irrelevant motor actions and grasp types as affordance labels and introducing the concept of mechanical action to represent the action possibilities between two objects. We provide new annotations by applying this scheme to the EPIC-KITCHENS dataset and test our annotation with tasks such as affordance recognition, hand-object interaction hotspots prediction, and cross-domain evaluation of affordance. The results show that models trained with our annotation can distinguish affordance from other concepts, predict fine-grained interaction possibilities on objects, and generalize through different domains.
comment: WACV. arXiv admin note: substantial text overlap with arXiv:2206.05424
☆ Improving CT Image Segmentation Accuracy Using StyleGAN Driven Data Augmentation
Medical Image Segmentation is a useful application for medical image analysis including detecting diseases and abnormalities in imaging modalities such as MRI, CT etc. Deep learning has proven to be promising for this task but usually has a low accuracy because of the lack of appropriate publicly available annotated or segmented medical datasets. In addition, the datasets that are available may have a different texture because of different dosage values or scanner properties than the images that need to be segmented. This paper presents a StyleGAN-driven approach for segmenting publicly available large medical datasets by using readily available extremely small annotated datasets in similar modalities. The approach involves augmenting the small segmented dataset and eliminating texture differences between the two datasets. The dataset is augmented by being passed through six different StyleGANs that are trained on six different style images taken from the large non-annotated dataset we want to segment. Specifically, style transfer is used to augment the training dataset. The annotations of the training dataset are hence combined with the textures of the non-annotated dataset to generate new anatomically sound images. The augmented dataset is then used to train a U-Net segmentation network which displays a significant improvement in the segmentation accuracy in segmenting the large non-annotated dataset.
comment: 17th International Meeting on Fully Three-Dimensional Image Reconstruction in Radiology and Nuclear Medicine(Fully3D Conference)
☆ An End-to-End Two-Phase Deep Learning-Based workflow to Segment Man-made Objects Around Reservoirs
Reservoirs are fundamental infrastructures for the management of water resources. Constructions around them can negatively impact their quality. Such unauthorized constructions can be monitored by land cover mapping (LCM) remote sensing (RS) images. In this paper, we develop a new approach based on DL and image processing techniques for man-made object segmentation around the reservoirs. In order to segment man-made objects around the reservoirs in an end-to-end procedure, segmenting reservoirs and identifying the region of interest (RoI) around them are essential. In the proposed two-phase workflow, the reservoir is initially segmented using a DL model. A post-processing stage is proposed to remove errors such as floating vegetation. Next, the RoI around the reservoir (RoIaR) is identified using the proposed image processing techniques. Finally, the man-made objects in the RoIaR are segmented using a DL architecture. We trained the proposed workflow using collected Google Earth (GE) images of eight reservoirs in Brazil over two different years. The U-Net-based and SegNet-based architectures are trained to segment the reservoirs. To segment man-made objects in the RoIaR, we trained and evaluated four possible architectures, U-Net, FPN, LinkNet, and PSPNet. Although the collected data has a high diversity (for example, they belong to different states, seasons, resolutions, etc.), we achieved good performances in both phases. Furthermore, applying the proposed post-processing to the output of reservoir segmentation improves the precision in all studied reservoirs except two cases. We validated the prepared workflow with a reservoir dataset outside the training reservoirs. The results show high generalization ability of the prepared workflow.
comment: 21 pages, 13 figures
☆ Delving Deep into Simplicity Bias for Long-Tailed Image Recognition
Simplicity Bias (SB) is a phenomenon that deep neural networks tend to rely favorably on simpler predictive patterns but ignore some complex features when applied to supervised discriminative tasks. In this work, we investigate SB in long-tailed image recognition and find the tail classes suffer more severely from SB, which harms the generalization performance of such underrepresented classes. We empirically report that self-supervised learning (SSL) can mitigate SB and perform in complementary to the supervised counterpart by enriching the features extracted from tail samples and consequently taking better advantage of such rare samples. However, standard SSL methods are designed without explicitly considering the inherent data distribution in terms of classes and may not be optimal for long-tailed distributed data. To address this limitation, we propose a novel SSL method tailored to imbalanced data. It leverages SSL by triple diverse levels, i.e., holistic-, partial-, and augmented-level, to enhance the learning of predictive complex patterns, which provides the potential to overcome the severe SB on tail data. Both quantitative and qualitative experimental results on five long-tailed benchmark datasets show our method can effectively mitigate SB and significantly outperform the competing state-of-the-arts.
☆ Online Misinformation Video Detection: A Survey
With information consumption via online video streaming becoming increasingly popular, misinformation video poses a new threat to the health of the online information ecosystem. Though previous studies have made much progress in detecting misinformation in text and image formats, video-based misinformation brings new and unique challenges to automatic detection systems: 1) high information heterogeneity brought by various modalities, 2) blurred distinction between misleading video manipulation and ubiquitous artistic video editing, and 3) new patterns of misinformation propagation due to the dominant role of recommendation systems on online video platforms. To facilitate research on this challenging task, we conduct this survey to present advances in misinformation video detection research. We first analyze and characterize the misinformation video from three levels including signals, semantics, and intents. Based on the characterization, we systematically review existing works for detection from features of various modalities to techniques for clue integration. We also introduce existing resources including representative datasets and widely used tools. Besides summarizing existing studies, we discuss related areas and outline open issues and future directions to encourage and guide more research on misinformation video detection. Our corresponding public repository is available at https://github.com/ICTMCG/Awesome-Misinfo-Video-Detection.
comment: 10 pages, 2 figures
☆ Scaling Self-Supervised End-to-End Driving with Multi-View Attention Learning
On end-to-end driving, a large amount of expert driving demonstrations is used to train an agent that mimics the expert by predicting its control actions. This process is self-supervised on vehicle signals (e.g., steering angle, acceleration) and does not require extra costly supervision (human labeling). Yet, the improvement of existing self-supervised end-to-end driving models has mostly given room to modular end-to-end models where labeling data intensive format such as semantic segmentation are required during training time. However, we argue that the latest self-supervised end-to-end models were developed in sub-optimal conditions with low-resolution images and no attention mechanisms. Further, those models are confined with limited field of view and far from the human visual cognition which can quickly attend far-apart scene features, a trait that provides an useful inductive bias. In this context, we present a new end-to-end model, trained by self-supervised imitation learning, leveraging a large field of view and a self-attention mechanism. These settings are more contributing to the agent's understanding of the driving scene, which brings a better imitation of human drivers. With only self-supervised training data, our model yields almost expert performance in CARLA's Nocrash metrics and could be rival to the SOTA models requiring large amounts of human labeled data. To facilitate further research, our code will be released.
☆ On the Ideal Number of Groups for Isometric Gradient Propagation
Recently, various normalization layers have been proposed to stabilize the training of deep neural networks. Among them, group normalization is a generalization of layer normalization and instance normalization by allowing a degree of freedom in the number of groups it uses. However, to determine the optimal number of groups, trial-and-error-based hyperparameter tuning is required, and such experiments are time-consuming. In this study, we discuss a reasonable method for setting the number of groups. First, we find that the number of groups influences the gradient behavior of the group normalization layer. Based on this observation, we derive the ideal number of groups, which calibrates the gradient scale to facilitate gradient descent optimization. Our proposed number of groups is theoretically grounded, architecture-aware, and can provide a proper value in a layer-wise manner for all layers. The proposed method exhibited improved performance over existing methods in numerous neural network architectures, tasks, and datasets.
comment: 10 pages, 2 figures
♻ ☆ Supervised Metric Learning to Rank for Retrieval via Contextual Similarity Optimization
There is extensive interest in metric learning methods for image retrieval. Many metric learning loss functions focus on learning a correct ranking of training samples, but strongly overfit semantically inconsistent labels and require a large amount of data. To address these shortcomings, we propose a new metric learning method, called contextual loss, which optimizes contextual similarity in addition to cosine similarity. Our contextual loss implicitly enforces semantic consistency among neighbors while converging to the correct ranking. We empirically show that the proposed loss is more robust to label noise, and is less prone to overfitting even when a large portion of train data is withheld. Extensive experiments demonstrate that our method achieves a new state-of-the-art across four image retrieval benchmarks and multiple different evaluation settings. Code is available at: https://github.com/Chris210634/metric-learning-using-contextual-similarity
♻ ☆ AV-NeRF: Learning Neural Fields for Real-World Audio-Visual Scene Synthesis
Human perception of the complex world relies on a comprehensive analysis of multi-modal signals, and the co-occurrences of audio and video signals provide humans with rich cues. This paper focuses on novel audio-visual scene synthesis in the real world. Given a video recording of an audio-visual scene, the task is to synthesize new videos with spatial audios along arbitrary novel camera trajectories in that audio-visual scene. Directly using a NeRF-based model for audio synthesis is insufficient due to its lack of prior knowledge and acoustic supervision. To tackle the challenges, we first propose an acoustic-aware audio generation module that integrates our prior knowledge of audio propagation into NeRF, in which we associate audio generation with the 3D geometry of the visual environment. In addition, we propose a coordinate transformation module that expresses a viewing direction relative to the sound source. Such a direction transformation helps the model learn sound source-centric acoustic fields. Moreover, we utilize a head-related impulse response function to synthesize pseudo binaural audio for data augmentation that strengthens training. We qualitatively and quantitatively demonstrate the advantage of our model on real-world audio-visual scenes. We refer interested readers to view our video results for convincing comparisons.
♻ ☆ Unsupervised Scene Adaptation with Memory Regularization in vivo IJCAI 2020
We consider the unsupervised scene adaptation problem of learning from both labeled source data and unlabeled target data. Existing methods focus on minoring the inter-domain gap between the source and target domains. However, the intra-domain knowledge and inherent uncertainty learned by the network are under-explored. In this paper, we propose an orthogonal method, called memory regularization in vivo to exploit the intra-domain knowledge and regularize the model training. Specifically, we refer to the segmentation model itself as the memory module, and minor the discrepancy of the two classifiers, i.e., the primary classifier and the auxiliary classifier, to reduce the prediction inconsistency. Without extra parameters, the proposed method is complementary to the most existing domain adaptation methods and could generally improve the performance of existing methods. Albeit simple, we verify the effectiveness of memory regularization on two synthetic-to-real benchmarks: GTA5 -> Cityscapes and SYNTHIA -> Cityscapes, yielding +11.1% and +11.3% mIoU improvement over the baseline model, respectively. Besides, a similar +12.0% mIoU improvement is observed on the cross-city benchmark: Cityscapes -> Oxford RobotCar.
comment: 7 pages, 4 figures, 6 tables (accepted by IJCAI 2020)
♻ ☆ OTRE: Where Optimal Transport Guided Unpaired Image-to-Image Translation Meets Regularization by Enhancing
Non-mydriatic retinal color fundus photography (CFP) is widely available due to the advantage of not requiring pupillary dilation, however, is prone to poor quality due to operators, systemic imperfections, or patient-related causes. Optimal retinal image quality is mandated for accurate medical diagnoses and automated analyses. Herein, we leveraged the Optimal Transport (OT) theory to propose an unpaired image-to-image translation scheme for mapping low-quality retinal CFPs to high-quality counterparts. Furthermore, to improve the flexibility, robustness, and applicability of our image enhancement pipeline in the clinical practice, we generalized a state-of-the-art model-based image reconstruction method, regularization by denoising, by plugging in priors learned by our OT-guided image-to-image translation network. We named it as regularization by enhancing (RE). We validated the integrated framework, OTRE, on three publicly available retinal image datasets by assessing the quality after enhancement and their performance on various downstream tasks, including diabetic retinopathy grading, vessel segmentation, and diabetic lesion segmentation. The experimental results demonstrated the superiority of our proposed framework over some state-of-the-art unsupervised competitors and a state-of-the-art supervised method.
comment: Accepted as a conference paper to The 28th biennial international conference on Information Processing in Medical Imaging (IPMI 2023)
♻ ☆ Take One Gram of Neural Features, Get Enhanced Group Robustness ECCV 2022
Predictive performance of machine learning models trained with empirical risk minimization (ERM) can degrade considerably under distribution shifts. The presence of spurious correlations in training datasets leads ERM-trained models to display high loss when evaluated on minority groups not presenting such correlations. Extensive attempts have been made to develop methods improving worst-group robustness. However, they require group information for each training input or at least, a validation set with group labels to tune their hyperparameters, which may be expensive to get or unknown a priori. In this paper, we address the challenge of improving group robustness without group annotation during training or validation. To this end, we propose to partition the training dataset into groups based on Gram matrices of features extracted by an ``identification'' model and to apply robust optimization based on these pseudo-groups. In the realistic context where no group labels are available, our experiments show that our approach not only improves group robustness over ERM but also outperforms all recent baselines
comment: Long version (Previous version: OOD-CV Workshop @ ECCV 2022)
♻ ☆ Medical Image Understanding with Pretrained Vision Language Models: A Comprehensive Study ICLR2023
The large-scale pre-trained vision language models (VLM) have shown remarkable domain transfer capability on natural images. However, it remains unknown whether this capability can also apply to the medical image domain. This paper thoroughly studies the knowledge transferability of pre-trained VLMs to the medical domain, where we show that well-designed medical prompts are the key to elicit knowledge from pre-trained VLMs. We demonstrate that by prompting with expressive attributes that are shared between domains, the VLM can carry the knowledge across domains and improve its generalization. This mechanism empowers VLMs to recognize novel objects with fewer or without image samples. Furthermore, to avoid the laborious manual designing process, we develop three approaches for automatic generation of medical prompts, which can inject expert-level medical knowledge and image-specific information into the prompts for fine-grained grounding. We conduct extensive experiments on thirteen different medical datasets across various modalities, showing that our well-designed prompts greatly improve the zero-shot performance compared to the default prompts, and our fine-tuned models surpass the supervised models by a significant margin.
comment: Accepted to ICLR2023
♻ ☆ Zero-shot Active Visual Search (ZAVIS): Intelligent Object Search for Robotic Assistants ICRA 2023
In this paper, we focus on the problem of efficiently locating a target object described with free-form language using a mobile robot equipped with vision sensors (e.g., an RGBD camera). Conventional active visual search predefines a set of objects to search for, rendering these techniques restrictive in practice. To provide added flexibility in active visual searching, we propose a system where a user can enter target commands using free-form language; we call this system Active Visual Search in the Wild (AVSW). AVSW detects and plans to search for a target object inputted by a user through a semantic grid map represented by static landmarks (e.g., desk or bed). For efficient planning of object search patterns, AVSW considers commonsense knowledge-based co-occurrence and predictive uncertainty while deciding which landmarks to visit first. We validate the proposed method with respect to SR (success rate) and SPL (success weighted by path length) in both simulated and real-world environments. The proposed method outperforms previous methods in terms of SPL in simulated scenarios with an average gap of 0.283. We further demonstrate AVSW with a Pioneer-3AT robot in real-world studies.
comment: To be appear on ICRA 2023
♻ ☆ Learning What and Where: Disentangling Location and Identity Tracking Without Supervision ICLR 2023
Our brain can almost effortlessly decompose visual data streams into background and salient objects. Moreover, it can anticipate object motion and interactions, which are crucial abilities for conceptual planning and reasoning. Recent object reasoning datasets, such as CATER, have revealed fundamental shortcomings of current vision-based AI systems, particularly when targeting explicit object representations, object permanence, and object reasoning. Here we introduce a self-supervised LOCation and Identity tracking system (Loci), which excels on the CATER tracking challenge. Inspired by the dorsal and ventral pathways in the brain, Loci tackles the binding problem by processing separate, slot-wise encodings of `what' and `where'. Loci's predictive coding-like processing encourages active error minimization, such that individual slots tend to encode individual objects. Interactions between objects and object dynamics are processed in the disentangled latent space. Truncated backpropagation through time combined with forward eligibility accumulation significantly speeds up learning and improves memory efficiency. Besides exhibiting superior performance in current benchmarks, Loci effectively extracts objects from video streams and separates them into location and Gestalt components. We believe that this separation offers a representation that will facilitate effective planning and reasoning on conceptual levels.
comment: Accepted at ICLR 2023
♻ ☆ FastPillars: A Deployment-friendly Pillar-based 3D Detector
The deployment of 3D detectors strikes one of the major challenges in real-world self-driving scenarios. Existing BEV-based (i.e., Bird Eye View) detectors favor sparse convolution (known as SPConv) to speed up training and inference, which puts a hard barrier for deployment especially for on-device applications. In this paper, we tackle the problem of efficient 3D object detection from LiDAR point clouds with deployment in mind. To reduce computational burden, we propose a pillar-based 3D detector with high performance from an industry perspective, termed FastPillars. Compared with previous methods, we introduce a more effective Max-and-Attention pillar encoding (MAPE) module, and redesigning a powerful and lightweight backbone CRVNet imbued with Cross Stage Partial network (CSP) in a reparameterization style, forming a compact feature representation framework. Extensive experiments demonstrate that our FastPillars surpasses the state-of-the-art 3D detectors regarding both on-device speed and performance. Specifically, FastPillars can be effectively deployed through TensorRT, obtaining real-time performance (24FPS) on a single RTX3070Ti GPU with 64.6 mAP on the nuScenes test set. Our code is publicly available at: https://github.com/StiphyJay/FastPillars.
♻ ☆ PatchDCT: Patch Refinement for High Quality Instance Segmentation ICLR 2023
High-quality instance segmentation has shown emerging importance in computer vision. Without any refinement, DCT-Mask directly generates high-resolution masks by compressed vectors. To further refine masks obtained by compressed vectors, we propose for the first time a compressed vector based multi-stage refinement framework. However, the vanilla combination does not bring significant gains, because changes in some elements of the DCT vector will affect the prediction of the entire mask. Thus, we propose a simple and novel method named PatchDCT, which separates the mask decoded from a DCT vector into several patches and refines each patch by the designed classifier and regressor. Specifically, the classifier is used to distinguish mixed patches from all patches, and to correct previously mispredicted foreground and background patches. In contrast, the regressor is used for DCT vector prediction of mixed patches, further refining the segmentation quality at boundary locations. Experiments on COCO show that our method achieves 2.0%, 3.2%, 4.5% AP and 3.4%, 5.3%, 7.0% Boundary AP improvements over Mask-RCNN on COCO, LVIS, and Cityscapes, respectively. It also surpasses DCT-Mask by 0.7%, 1.1%, 1.3% AP and 0.9%, 1.7%, 4.2% Boundary AP on COCO, LVIS and Cityscapes. Besides, the performance of PatchDCT is also competitive with other state-of-the-art methods.
comment: 15 pages, 7 figures, 13 tables, accepted by ICLR 2023, the source code is available at https://github.com/olivia-w12/PatchDCT
♻ ☆ Name Your Colour For the Task: Artificially Discover Colour Naming via Colour Quantisation Transformer
The long-standing theory that a colour-naming system evolves under dual pressure of efficient communication and perceptual mechanism is supported by more and more linguistic studies, including analysing four decades of diachronic data from the Nafaanra language. This inspires us to explore whether machine learning could evolve and discover a similar colour-naming system via optimising the communication efficiency represented by high-level recognition performance. Here, we propose a novel colour quantisation transformer, CQFormer, that quantises colour space while maintaining the accuracy of machine recognition on the quantised images. Given an RGB image, Annotation Branch maps it into an index map before generating the quantised image with a colour palette; meanwhile the Palette Branch utilises a key-point detection way to find proper colours in the palette among the whole colour space. By interacting with colour annotation, CQFormer is able to balance both the machine vision accuracy and colour perceptual structure such as distinct and stable colour distribution for discovered colour system. Very interestingly, we even observe the consistent evolution pattern between our artificial colour system and basic colour terms across human languages. Besides, our colour quantisation method also offers an efficient quantisation method that effectively compresses the image storage while maintaining high performance in high-level recognition tasks such as classification and detection. Extensive experiments demonstrate the superior performance of our method with extremely low bit-rate colours, showing potential to integrate into quantisation network to quantities from image to network activation. We will release the source code upon acceptance.
♻ ☆ Diffusion Denoising for Low-Dose-CT Model
Low-dose Computed Tomography (LDCT) reconstruction is an important task in medical image analysis. Recent years have seen many deep learning based methods, proved to be effective in this area. However, these methods mostly follow a supervised architecture, which needs paired CT image of full dose and quarter dose, and the solution is highly dependent on specific measurements. In this work, we introduce Denoising Diffusion LDCT Model, dubbed as DDLM, generating noise-free CT image using conditioned sampling. DDLM uses pretrained model, and need no training nor tuning process, thus our proposal is in unsupervised manner. Experiments on LDCT images have shown comparable performance of DDLM using less inference time, surpassing other state-of-the-art methods, proving both accurate and efficient. Implementation code will be set to public soon.
comment: The method and experiment of this paper has some error, and we need to revise it
♻ ☆ Lightweight Event-based Optical Flow Estimation via Iterative Deblurring
Inspired by frame-based methods, state-of-the-art event-based optical flow networks rely on the explicit computation of correlation volumes, which are expensive to compute and store on systems with limited processing budget and memory. To this end, we introduce IDNet (Iterative Deblurring Network), a lightweight yet well-performing event-based optical flow network without using correlation volumes. IDNet leverages the unique spatiotemporally continuous nature of event streams to propose an alternative way of implicitly capturing correlation through iterative refinement and motion deblurring. Our network does not compute correlation volumes but rather utilizes a recurrent network to maximize the spatiotemporal correlation of events iteratively. We further propose two iterative update schemes: "ID" which iterates over the same batch of events, and "TID" which iterates over time with streaming events in an online fashion. Benchmark results show the former "ID" scheme can reach close to state-of-the-art performance with 33% of savings in compute and 90% in memory footprint, while the latter "TID" scheme is even more efficient promising 83% of compute savings and 15 times less latency at the cost of 18% of performance drop.
comment: Added supplementary materials
♻ ☆ Rethinking Persistent Homology for Visual Recognition ICML 2022
Persistent topological properties of an image serve as an additional descriptor providing an insight that might not be discovered by traditional neural networks. The existing research in this area focuses primarily on efficiently integrating topological properties of the data in the learning process in order to enhance the performance. However, there is no existing study to demonstrate all possible scenarios where introducing topological properties can boost or harm the performance. This paper performs a detailed analysis of the effectiveness of topological properties for image classification in various training scenarios, defined by: the number of training samples, the complexity of the training data and the complexity of the backbone network. We identify the scenarios that benefit the most from topological features, e.g., training simple networks on small datasets. Additionally, we discuss the problem of topological consistency of the datasets which is one of the major bottlenecks for using topological features for classification. We further demonstrate how the topological inconsistency can harm the performance for certain scenarios.
comment: ICML 2022 Workshop on Topology, Algebra, and Geometry in Machine Learning
♻ ☆ Novel Fundus Image Preprocessing for Retcam Images to Improve Deep Learning Classification of Retinopathy of Prematurity
Retinopathy of Prematurity (ROP) is a potentially blinding eye disorder because of damage to the eye's retina which can affect babies born prematurely. Screening of ROP is essential for early detection and treatment. This is a laborious and manual process which requires trained physician performing dilated ophthalmological examination which can be subjective resulting in lower diagnosis success for clinically significant disease. Automated diagnostic methods can assist ophthalmologists increase diagnosis accuracy using deep learning. Several research groups have highlighted various approaches. This paper proposes the use of new novel fundus preprocessing methods using pretrained transfer learning frameworks to create hybrid models to give higher diagnosis accuracy. The evaluations show that these novel methods in comparison to traditional imaging processing contribute to higher accuracy in classifying Plus disease, Stages of ROP and Zones. We achieve accuracy of 97.65% for Plus disease, 89.44% for Stage, 90.24% for Zones with limited training dataset.
comment: 10 pages, 4 figures, 7 tables. arXiv admin note: text overlap with arXiv:1904.08796 by other authors
♻ ☆ FG-Depth: Flow-Guided Unsupervised Monocular Depth Estimation ICRA2023
The great potential of unsupervised monocular depth estimation has been demonstrated by many works due to low annotation cost and impressive accuracy comparable to supervised methods. To further improve the performance, recent works mainly focus on designing more complex network structures and exploiting extra supervised information, e.g., semantic segmentation. These methods optimize the models by exploiting the reconstructed relationship between the target and reference images in varying degrees. However, previous methods prove that this image reconstruction optimization is prone to get trapped in local minima. In this paper, our core idea is to guide the optimization with prior knowledge from pretrained Flow-Net. And we show that the bottleneck of unsupervised monocular depth estimation can be broken with our simple but effective framework named FG-Depth. In particular, we propose (i) a flow distillation loss to replace the typical photometric loss that limits the capacity of the model and (ii) a prior flow based mask to remove invalid pixels that bring the noise in training loss. Extensive experiments demonstrate the effectiveness of each component, and our approach achieves state-of-the-art results on both KITTI and NYU-Depth-v2 datasets.
comment: Accepted by ICRA2023
♻ ☆ Video compression dataset and benchmark of learning-based video-quality metrics
Video-quality measurement is a critical task in video processing. Nowadays, many implementations of new encoding standards - such as AV1, VVC, and LCEVC - use deep-learning-based decoding algorithms with perceptual metrics that serve as optimization objectives. But investigations of the performance of modern video- and image-quality metrics commonly employ videos compressed using older standards, such as AVC. In this paper, we present a new benchmark for video-quality metrics that evaluates video compression. It is based on a new dataset consisting of about 2,500 streams encoded using different standards, including AVC, HEVC, AV1, VP9, and VVC. Subjective scores were collected using crowdsourced pairwise comparisons. The list of evaluated metrics includes recent ones based on machine learning and neural networks. The results demonstrate that new no-reference metrics exhibit a high correlation with subjective quality and approach the capability of top full-reference metrics.
comment: 10 pages, 4 figures, 6 tables, 1 supplementary material
♻ ☆ Single Shot Self-Reliant Scene Text Spotter by Decoupled yet Collaborative Detection and Recognition
Typical text spotters follow the two-stage spotting paradigm which detects the boundary for a text instance first and then performs text recognition within the detected regions. Despite the remarkable progress of such spotting paradigm, an important limitation is that the performance of text recognition depends heavily on the precision of text detection, resulting in the potential error propagation from detection to recognition. In this work, we propose the single shot Self-Reliant Scene Text Spotter v2 (SRSTS v2), which circumvents this limitation by decoupling recognition from detection while optimizing two tasks collaboratively. Specifically, our SRSTS v2 samples representative feature points around each potential text instance, and conducts both text detection and recognition in parallel guided by these sampled points. Thus, the text recognition is no longer dependent on detection, thereby alleviating the error propagation from detection to recognition. Moreover, the sampling module is learned under the supervision from both detection and recognition, which allows for the collaborative optimization and mutual enhancement between two tasks. Benefiting from such sampling-driven concurrent spotting framework, our approach is able to recognize the text instances correctly even if the precise text boundaries are challenging to detect. Extensive experiments on four benchmarks demonstrate that our method compares favorably to state-of-the-art spotters.
♻ ☆ Flattening-Net: Deep Regular 2D Representation for 3D Point Cloud Analysis
Point clouds are characterized by irregularity and unstructuredness, which pose challenges in efficient data exploitation and discriminative feature extraction. In this paper, we present an unsupervised deep neural architecture called Flattening-Net to represent irregular 3D point clouds of arbitrary geometry and topology as a completely regular 2D point geometry image (PGI) structure, in which coordinates of spatial points are captured in colors of image pixels. \mr{Intuitively, Flattening-Net implicitly approximates a locally smooth 3D-to-2D surface flattening process while effectively preserving neighborhood consistency.} \mr{As a generic representation modality, PGI inherently encodes the intrinsic property of the underlying manifold structure and facilitates surface-style point feature aggregation.} To demonstrate its potential, we construct a unified learning framework directly operating on PGIs to achieve \mr{diverse types of high-level and low-level} downstream applications driven by specific task networks, including classification, segmentation, reconstruction, and upsampling. Extensive experiments demonstrate that our methods perform favorably against the current state-of-the-art competitors. We will make the code and data publicly available at https://github.com/keeganhk/Flattening-Net.
comment: Accepted to TPAMI
♻ ☆ R2FD2: Fast and Robust Matching of Multimodal Remote Sensing Image via Repeatable Feature Detector and Rotation-invariant Feature Descriptor
Automatically identifying feature correspondences between multimodal images is facing enormous challenges because of the significant differences both in radiation and geometry. To address these problems, we propose a novel feature matching method (named R2FD2) that is robust to radiation and rotation differences. Our R2FD2 is conducted in two critical contributions, consisting of a repeatable feature detector and a rotation-invariant feature descriptor. In the first stage, a repeatable feature detector called the Multi-channel Auto-correlation of the Log-Gabor (MALG) is presented for feature detection, which combines the multi-channel auto-correlation strategy with the Log-Gabor wavelets to detect interest points (IPs) with high repeatability and uniform distribution. In the second stage, a rotation-invariant feature descriptor is constructed, named the Rotation-invariant Maximum index map of the Log-Gabor (RMLG), which consists of two components: fast assignment of dominant orientation and construction of feature representation. In the process of fast assignment of dominant orientation, a Rotation-invariant Maximum Index Map (RMIM) is built to address rotation deformations. Then, the proposed RMLG incorporates the rotation-invariant RMIM with the spatial configuration of DAISY to depict a more discriminative feature representation, which improves RMLG's resistance to radiation and rotation variances.Experimental results show that the proposed R2FD2 outperforms five state-of-the-art feature matching methods, and has superior advantages in adaptability and universality. Moreover, our R2FD2 achieves the accuracy of matching within two pixels and has a great advantage in matching efficiency over other state-of-the-art methods.
comment: 33 pages, 15 figures
♻ ☆ Advances and Challenges in Multimodal Remote Sensing Image Registration
Over the past few decades, with the rapid development of global aerospace and aerial remote sensing technology, the types of sensors have evolved from the traditional monomodal sensors (e.g., optical sensors) to the new generation of multimodal sensors [e.g., multispectral, hyperspectral, light detection and ranging (LiDAR) and synthetic aperture radar (SAR) sensors]. These advanced devices can dynamically provide various and abundant multimodal remote sensing images with different spatial, temporal, and spectral resolutions according to different application requirements. Since then, it is of great scientific significance to carry out the research of multimodal remote sensing image registration, which is a crucial step for integrating the complementary information among multimodal data and making comprehensive observations and analysis of the Earths surface. In this work, we will present our own contributions to the field of multimodal image registration, summarize the advantages and limitations of existing multimodal image registration methods, and then discuss the remaining challenges and make a forward-looking prospect for the future development of the field.
comment: 10 pages, 4 figures
♻ ☆ Revitalize Region Feature for Democratizing Video-Language Pre-training of Retrieval
Recent dominant methods for video-language pre-training (VLP) learn transferable representations from the raw pixels in an end-to-end manner to achieve advanced performance on downstream video-language retrieval. Despite the impressive results, VLP research becomes extremely expensive with the need for massive data and a long training time, preventing further explorations. In this work, we revitalize region features of sparsely sampled video clips to significantly reduce both spatial and temporal visual redundancy towards democratizing VLP research at the same time achieving state-of-the-art results. Specifically, to fully explore the potential of region features, we introduce a novel bidirectional region-word alignment regularization that properly optimizes the fine-grained relations between regions and certain words in sentences, eliminating the domain/modality disconnections between pre-extracted region features and text. Extensive results of downstream video-language retrieval tasks on four datasets demonstrate the superiority of our method on both effectiveness and efficiency, \textit{e.g.}, our method achieves competing results with 80\% fewer data and 85\% less pre-training time compared to the most efficient VLP method so far \cite{lei2021less}. The code will be available at \url{https://github.com/showlab/DemoVLP}.
♻ ☆ A Survey of Supernet Optimization and its Applications: Spatial and Temporal Optimization for Neural Architecture Search
This survey focuses on categorizing and evaluating the methods of supernet optimization in the field of Neural Architecture Search (NAS). Supernet optimization involves training a single, over-parameterized network that encompasses the search space of all possible network architectures. The survey analyses supernet optimization methods based on their approaches to spatial and temporal optimization. Spatial optimization relates to optimizing the architecture and parameters of the supernet and its subnets, while temporal optimization deals with improving the efficiency of selecting architectures from the supernet. The benefits, limitations, and potential applications of these methods in various tasks and settings, including transferability, domain generalization, and Transformer models, are also discussed.
♻ ☆ Learning Continuous Rotation Canonicalization with Radial Beam Sampling
Nearly all state of the art vision models are sensitive to image rotations. Existing methods often compensate for missing inductive biases by using augmented training data to learn pseudo-invariances. Alongside the resource demanding data inflation process, predictions often poorly generalize. The inductive biases inherent to convolutional neural networks allow for translation equivariance through kernels acting parallely to the horizontal and vertical axes of the pixel grid. This inductive bias, however, does not allow for rotation equivariance. We propose a radial beam sampling strategy along with radial kernels operating on these beams to inherently incorporate center-rotation covariance. Together with an angle distance loss, we present a radial beam-based image canonicalization model, short BIC. Our model allows for maximal continuous angle regression and canonicalizes arbitrary center-rotated input images. As a pre-processing model, this enables rotation-invariant vision pipelines with model-agnostic rotation-sensitive downstream predictions. We show that our end-to-end trained angle regressor is able to predict continuous rotation angles on several vision datasets, i.e. FashionMNIST, CIFAR10, COIL100, and LFW.
♻ ☆ Compound Prototype Matching for Few-shot Action Recognition ECCV 2022
Few-shot action recognition aims to recognize novel action classes using only a small number of labeled training samples. In this work, we propose a novel approach that first summarizes each video into compound prototypes consisting of a group of global prototypes and a group of focused prototypes, and then compares video similarity based on the prototypes. Each global prototype is encouraged to summarize a specific aspect from the entire video, for example, the start/evolution of the action. Since no clear annotation is provided for the global prototypes, we use a group of focused prototypes to focus on certain timestamps in the video. We compare video similarity by matching the compound prototypes between the support and query videos. The global prototypes are directly matched to compare videos from the same perspective, for example, to compare whether two actions start similarly. For the focused prototypes, since actions have various temporal variations in the videos, we apply bipartite matching to allow the comparison of actions with different temporal positions and shifts. Experiments demonstrate that our proposed method achieves state-of-the-art results on multiple benchmarks.
comment: ECCV 2022
♻ ☆ Rethinking Semi-Supervised Medical Image Segmentation: A Variance-Reduction Perspective
For medical image segmentation, contrastive learning is the dominant practice to improve the quality of visual representations by contrasting semantically similar and dissimilar pairs of samples. This is enabled by the observation that without accessing ground truth label, negative examples with truly dissimilar anatomical features, if sampled, can significantly improve the performance. In reality, however, these samples may come from similar anatomical features and the models may struggle to distinguish the minority tail-class samples, making the tail classes more prone to misclassification, both of which typically lead to model collapse. In this paper, we propose ARCO, a semi-supervised contrastive learning (CL) framework with stratified group sampling theory in medical image segmentation. In particular, we first propose building ARCO through the concept of variance-reduced estimation, and show that certain variance-reduction techniques are particularly beneficial in medical image segmentation tasks with extremely limited labels. Furthermore, we theoretically prove these sampling techniques are universal in variance reduction. Finally, we experimentally validate our approaches on three benchmark datasets with different label settings, and our methods consistently outperform state-of-the-art semi- and fully-supervised methods. Additionally, we augment the CL frameworks with these sampling techniques and demonstrate significant gains over previous methods. We believe our work is an important step towards semi-supervised medical image segmentation by quantifying the limitation of current self-supervision objectives for accomplishing medical image analysis tasks.
♻ ☆ Joint stereo 3D object detection and implicit surface reconstruction
We present a new learning-based framework S-3D-RCNN that can recover accurate object orientation in SO(3) and simultaneously predict implicit shapes for outdoor rigid objects from stereo RGB images. In contrast to previous studies that map local appearance to observation angles, we explore a progressive approach by extracting meaningful Intermediate Geometrical Representations (IGRs) to estimate egocentric object orientation. This approach features a deep model that transforms perceived intensities to object part coordinates, which are mapped to a 3D representation encoding object orientation in the camera coordinate system. To enable implicit shape estimation, the IGRs are further extended to model visible object surface with a point-based representation and explicitly addresses the unseen surface hallucination problem. Extensive experiments validate the effectiveness of the proposed IGRs and S-3D-RCNN achieves superior 3D scene understanding performance using existing and proposed new metrics on the KITTI benchmark. Code and pre-trained models will be available at this https URL.
♻ ☆ PriorLane: A Prior Knowledge Enhanced Lane Detection Approach Based on Transformer ICRA 2023
Lane detection is one of the fundamental modules in self-driving. In this paper we employ a transformer-only method for lane detection, thus it could benefit from the blooming development of fully vision transformer and achieve the state-of-the-art (SOTA) performance on both CULane and TuSimple benchmarks, by fine-tuning the weight fully pre-trained on large datasets. More importantly, this paper proposes a novel and general framework called PriorLane, which is used to enhance the segmentation performance of the fully vision transformer by introducing the low-cost local prior knowledge. Specifically, PriorLane utilizes an encoder-only transformer to fuse the feature extracted by a pre-trained segmentation model with prior knowledge embeddings. Note that a Knowledge Embedding Alignment (KEA) module is adapted to enhance the fusion performance by aligning the knowledge embedding. Extensive experiments on our Zjlab dataset show that PriorLane outperforms SOTA lane detection methods by a 2.82% mIoU when prior knowledge is employed, and the code will be released at: https://github.com/vincentqqb/PriorLane.
comment: Accepted by ICRA 2023
♻ ☆ CHiLS: Zero-Shot Image Classification with Hierarchical Label Sets
Open vocabulary models (e.g. CLIP) have shown strong performance on zero-shot classification through their ability generate embeddings for each class based on their (natural language) names. Prior work has focused on improving the accuracy of these models through prompt engineering or by incorporating a small amount of labeled downstream data (via finetuning). However, there has been little focus on improving the richness of the class names themselves, which can pose issues when class labels are coarsely-defined and uninformative. We propose Classification with Hierarchical Label Sets (or CHiLS), an alternative strategy for zero-shot classification specifically designed for datasets with implicit semantic hierarchies. CHiLS proceeds in three steps: (i) for each class, produce a set of subclasses, using either existing label hierarchies or by querying GPT-3; (ii) perform the standard zero-shot CLIP procedure as though these subclasses were the labels of interest; (iii) map the predicted subclass back to its parent to produce the final prediction. Across numerous datasets with underlying hierarchical structure, CHiLS leads to improved accuracy in situations both with and without ground-truth hierarchical information. CHiLS is simple to implement within existing CLIP pipelines and requires no additional training cost. Code is available at: https://github.com/acmi-lab/CHILS.
♻ ☆ Low-Light Image and Video Enhancement: A Comprehensive Survey and Beyond
This paper presents a comprehensive survey of low-light image and video enhancement. We begin with the challenging mixed over-/under-exposed images, which are under-performed by existing methods. To this end, we propose two variants of the SICE dataset named SICE\_Grad and SICE\_Mix. Next, we introduce Night Wenzhou, a large-scale, high-resolution video dataset, to address the lack of low-light video datasets that discourages the use of low-light image enhancement (LLIE) methods in videos. Our Night Wenzhou dataset is challenging since it consists of fast-moving aerial scenes and streetscapes with varying illuminations and degradation. We then construct a hierarchical taxonomy, conduct extensive key technique analysis, and performs experimental comparisons for representative LLIE approaches using our proposed datasets and the current benchmark datasets. Finally, we identify emerging applications, address unresolved challenges, and propose future research topics for the LLIE community. Our datasets are available at https://github.com/ShenZheng2000/LLIE_Survey.
comment: 21 pages, 9 tables, and 25 figures
♻ ☆ GEDI: GEnerative and DIscriminative Training for Self-Supervised Learning
Self-supervised learning is a popular and powerful method for utilizing large amounts of unlabeled data, for which a wide variety of training objectives have been proposed in the literature. In this study, we perform a Bayesian analysis of state-of-the-art self-supervised learning objectives and propose a unified formulation based on likelihood learning. Our analysis suggests a simple method for integrating self-supervised learning with generative models, allowing for the joint training of these two seemingly distinct approaches. We refer to this combined framework as GEDI, which stands for GEnerative and DIscriminative training. Additionally, we demonstrate an instantiation of the GEDI framework by integrating an energy-based model with a cluster-based self-supervised learning model. Through experiments on synthetic and real-world data, including SVHN, CIFAR10, and CIFAR100, we show that GEDI outperforms existing self-supervised learning strategies in terms of clustering performance by a wide margin. We also demonstrate that GEDI can be integrated into a neural-symbolic framework to address tasks in the small data regime, where it can use logical constraints to further improve clustering and classification performance.
comment: Fixed typos/cleaned the experimental section
♻ ☆ Improving Transfer Learning with a Dual Image and Video Transformer for Multi-label Movie Trailer Genre Classification
In this paper, we study the transferability of ImageNet spatial and Kinetics spatio-temporal representations to multi-label Movie Trailer Genre Classification (MTGC). In particular, we present an extensive evaluation of the transferability of ConvNet and Transformer models pretrained on ImageNet and Kinetics to Trailers12k, a new manually-curated movie trailer dataset composed of 12,000 videos labeled with 10 different genres and associated metadata. We analyze different aspects that can influence transferability, such as frame rate, input video extension, and spatio-temporal modeling. In order to reduce the spatio-temporal structure gap between ImageNet/Kinetics and Trailers12k, we propose Dual Image and Video Transformer Architecture (DIViTA), which performs shot detection so as to segment the trailer into highly correlated clips, providing a more cohesive input for pretrained backbones and improving transferability (a 1.83% increase for ImageNet and 3.75% for Kinetics). Our results demonstrate that representations learned on either ImageNet or Kinetics are comparatively transferable to Trailers12k. Moreover, both datasets provide complementary information that can be combined to improve classification performance (a 2.91% gain compared to the top single pretraining). Interestingly, using lightweight ConvNets as pretrained backbones resulted in only a 3.46% drop in classification performance compared with the top Transformer while requiring only 11.82% of its parameters and 0.81% of its FLOPS.
♻ ☆ Don't Blame the Annotator: Bias Already Starts in the Annotation Instructions EACL 2023
In recent years, progress in NLU has been driven by benchmarks. These benchmarks are typically collected by crowdsourcing, where annotators write examples based on annotation instructions crafted by dataset creators. In this work, we hypothesize that annotators pick up on patterns in the crowdsourcing instructions, which bias them to write many similar examples that are then over-represented in the collected data. We study this form of bias, termed instruction bias, in 14 recent NLU benchmarks, showing that instruction examples often exhibit concrete patterns, which are propagated by crowdworkers to the collected data. This extends previous work (Geva et al., 2019) and raises a new concern of whether we are modeling the dataset creator's instructions, rather than the task. Through a series of experiments, we show that, indeed, instruction bias can lead to overestimation of model performance, and that models struggle to generalize beyond biases originating in the crowdsourcing instructions. We further analyze the influence of instruction bias in terms of pattern frequency and model size, and derive concrete recommendations for creating future NLU benchmarks.
comment: Accepted to EACL 2023
♻ ☆ This Intestine Does Not Exist: Multiscale Residual Variational Autoencoder for Realistic Wireless Capsule Endoscopy Image Generation
Medical image synthesis has emerged as a promising solution to address the limited availability of annotated medical data needed for training machine learning algorithms in the context of image-based Clinical Decision Support (CDS) systems. To this end, Generative Adversarial Networks (GANs) have been mainly applied to support the algorithm training process by generating synthetic images for data augmentation. However, in the field of Wireless Capsule Endoscopy (WCE), the limited content diversity and size of existing publicly available annotated datasets, adversely affect both the training stability and synthesis performance of GANs. Aiming to a viable solution for WCE image synthesis, a novel Variational Autoencoder architecture is proposed, namely "This Intestine Does not Exist" (TIDE). The proposed architecture comprises multiscale feature extraction convolutional blocks and residual connections, which enable the generation of high-quality and diverse datasets even with a limited number of training images. Contrary to the current approaches, which are oriented towards the augmentation of the available datasets, this study demonstrates that using TIDE, real WCE datasets can be fully substituted by artificially generated ones, without compromising classification performance. Furthermore, qualitative and user evaluation studies by experienced WCE specialists, validate from a medical viewpoint that both the normal and abnormal WCE images synthesized by TIDE are sufficiently realistic.
comment: 10 pages
♻ ☆ Domain Re-Modulation for Few-Shot Generative Domain Adaptation
In this study, we investigate the task of few-shot Generative Domain Adaptation (GDA), which involves transferring a pre-trained generator from one domain to a new domain using one or a few reference images. Building upon previous research that has focused on Target-domain Consistency, Large Diversity, and Cross-domain Consistency, we conclude two additional desired properties for GDA: Memory and Domain Association. To meet these properties, we proposed a novel method Domain Re-Modulation (DoRM). Specifically, DoRM freezes the source generator and employs additional mapping and affine modules (M&A module) to capture the attributes of the target domain, resulting in a linearly combinable domain shift in style space. This allows for high-fidelity multi-domain and hybrid-domain generation by integrating multiple M&A modules in a single generator. DoRM is lightweight and easy to implement. Extensive experiments demonstrated the superior performance of DoRM on both one-shot and 10-shot GDA, both quantitatively and qualitatively. Additionally, for the first time, multi-domain and hybrid-domain generation can be achieved with a minimal storage cost by using a single model. The code will be available at https://github.com/wuyi2020/DoRM.
comment: Under Review
♻ ☆ NODEO: A Neural Ordinary Differential Equation Based Optimization Framework for Deformable Image Registration CVPR
Deformable image registration (DIR), aiming to find spatial correspondence between images, is one of the most critical problems in the domain of medical image analysis. In this paper, we present a novel, generic, and accurate diffeomorphic image registration framework that utilizes neural ordinary differential equations (NODEs). We model each voxel as a moving particle and consider the set of all voxels in a 3D image as a high-dimensional dynamical system whose trajectory determines the targeted deformation field. Our method leverages deep neural networks for their expressive power in modeling dynamical systems, and simultaneously optimizes for a dynamical system between the image pairs and the corresponding transformation. Our formulation allows various constraints to be imposed along the transformation to maintain desired regularities. Our experiment results show that our method outperforms the benchmarks under various metrics. Additionally, we demonstrate the feasibility to expand our framework to register multiple image sets using a unified form of transformation,which could possibly serve a wider range of applications.
comment: Accepted by the IEEE / CVF Computer Vision and Pattern Recognition Conference (CVPR) 2022
♻ ☆ Eliminating Prior Bias for Semantic Image Editing via Dual-Cycle Diffusion
The recent success of text-to-image generation diffusion models has also revolutionized semantic image editing, enabling the manipulation of images based on query/target texts. Despite these advancements, a significant challenge lies in the potential introduction of prior bias in pre-trained models during image editing, e.g., making unexpected modifications to inappropriate regions. To this point, we present a novel Dual-Cycle Diffusion model that addresses the issue of prior bias by generating an unbiased mask as the guidance of image editing. The proposed model incorporates a Bias Elimination Cycle that consists of both a forward path and an inverted path, each featuring a Structural Consistency Cycle to ensure the preservation of image content during the editing process. The forward path utilizes the pre-trained model to produce the edited image, while the inverted path converts the result back to the source image. The unbiased mask is generated by comparing differences between the processed source image and the edited image to ensure that both conform to the same distribution. Our experiments demonstrate the effectiveness of the proposed method, as it significantly improves the D-CLIP score from 0.272 to 0.283. The code will be available at https://github.com/JohnDreamer/DualCycleDiffsion.
♻ ☆ Forgetting to Remember: A Scalable Incremental Learning Framework for Cross-Task Blind Image Quality Assessment
Recent years have witnessed the great success of blind image quality assessment (BIQA) in various task-specific scenarios, which present invariable distortion types and evaluation criteria. However, due to the rigid structure and learning framework, they cannot apply to the cross-task BIQA scenario, where the distortion types and evaluation criteria keep changing in practical applications. This paper proposes a scalable incremental learning framework (SILF) that could sequentially conduct BIQA across multiple evaluation tasks with limited memory capacity. More specifically, we develop a dynamic parameter isolation strategy to sequentially update the task-specific parameter subsets, which are non-overlapped with each other. Each parameter subset is temporarily settled to Remember one evaluation preference toward its corresponding task, and the previously settled parameter subsets can be adaptively reused in the following BIQA to achieve better performance based on the task relevance. To suppress the unrestrained expansion of memory capacity in sequential tasks learning, we develop a scalable memory unit by gradually and selectively pruning unimportant neurons from previously settled parameter subsets, which enable us to Forget part of previous experiences and free the limited memory capacity for adapting to the emerging new tasks. Extensive experiments on eleven IQA datasets demonstrate that our proposed method significantly outperforms the other state-of-the-art methods in cross-task BIQA. The source code of the proposed method is available at https://github.com/maruiperfect/SILF.
♻ ☆ ShaSTA: Modeling Shape and Spatio-Temporal Affinities for 3D Multi-Object Tracking
Multi-object tracking is a cornerstone capability of any robotic system. The quality of tracking is largely dependent on the quality of the detector used. In many applications, such as autonomous vehicles, it is preferable to over-detect objects to avoid catastrophic outcomes due to missed detections. As a result, current state-of-the-art 3D detectors produce high rates of false-positives to ensure a low number of false-negatives. This can negatively affect tracking by making data association and track lifecycle management more challenging. Additionally, occasional false-negative detections due to difficult scenarios like occlusions can harm tracking performance. To address these issues in a unified framework, we propose to learn shape and spatio-temporal affinities between tracks and detections in consecutive frames. Our affinity provides a probabilistic matching that leads to robust data association, track lifecycle management, false-positive elimination, false-negative propagation, and sequential track confidence refinement. Though past 3D MOT approaches address a subset of components in this problem domain, we offer the first self-contained framework that addresses all these aspects of the 3D MOT problem. We quantitatively evaluate our method on the nuScenes tracking benchmark where we achieve 1st place amongst LiDAR-only trackers using CenterPoint detections. Our method estimates accurate and precise tracks, while decreasing the overall number of false-positive and false-negative tracks and increasing the number of true-positive tracks. We analyze our performance with 5 metrics, giving a comprehensive overview of our approach to indicate how our tracking framework may impact the ultimate goal of an autonomous mobile agent. We also present ablative experiments and qualitative results that demonstrate our framework's capabilities in complex scenarios.
comment: 10 pages, 3 figures
♻ ☆ Deep Learning for Diagonal Earlobe Crease Detection ICPR
An article published on Medical News Today in June 2022 presented a fundamental question in its title: Can an earlobe crease predict heart attacks? The author explained that end arteries supply the heart and ears. In other words, if they lose blood supply, no other arteries can take over, resulting in tissue damage. Consequently, some earlobes have a diagonal crease, line, or deep fold that resembles a wrinkle. In this paper, we take a step toward detecting this specific marker, commonly known as DELC or Frank's Sign. For this reason, we have made the first DELC dataset available to the public. In addition, we have investigated the performance of numerous cutting-edge backbones on annotated photos. Experimentally, we demonstrate that it is possible to solve this challenge by combining pre-trained encoders with a customized classifier to achieve 97.7% accuracy. Moreover, we have analyzed the backbone trade-off between performance and size, estimating MobileNet as the most promising encoder.
comment: Accepted at 12th International Conference on Pattern Recognition Applications (ICPRAM 2023)
♻ ☆ GOHSP: A Unified Framework of Graph and Optimization-based Heterogeneous Structured Pruning for Vision Transformer AAAI 2023
The recently proposed Vision transformers (ViTs) have shown very impressive empirical performance in various computer vision tasks, and they are viewed as an important type of foundation model. However, ViTs are typically constructed with large-scale sizes, which then severely hinder their potential deployment in many practical resources-constrained applications. To mitigate this challenging problem, structured pruning is a promising solution to compress model size and enable practical efficiency. However, unlike its current popularity for CNNs and RNNs, structured pruning for ViT models is little explored. In this paper, we propose GOHSP, a unified framework of Graph and Optimization-based Structured Pruning for ViT models. We first develop a graph-based ranking for measuring the importance of attention heads, and the extracted importance information is further integrated to an optimization-based procedure to impose the heterogeneous structured sparsity patterns on the ViT models. Experimental results show that our proposed GOHSP demonstrates excellent compression performance. On CIFAR-10 dataset, our approach can bring 40% parameters reduction with no accuracy loss for ViT-Small model. On ImageNet dataset, with 30% and 35% sparsity ratio for DeiT-Tiny and DeiT-Small models, our approach achieves 1.65% and 0.76% accuracy increase over the existing structured pruning methods, respectively.
comment: This manuscript was accepted to AAAI 2023 Main Track
Information Retrieval 12
☆ Optimizing Audio Recommendations for the Long-Term: A Reinforcement Learning Perspective
We study the problem of optimizing a recommender system for outcomes that occur over several weeks or months. We begin by drawing on reinforcement learning to formulate a comprehensive model of users' recurring relationships with a recommender system. Measurement, attribution, and coordination challenges complicate algorithm design. We describe careful modeling -- including a new representation of user state and key conditional independence assumptions -- which overcomes these challenges and leads to simple, testable recommender system prototypes. We apply our approach to a podcast recommender system that makes personalized recommendations to hundreds of millions of listeners. A/B tests demonstrate that purposefully optimizing for long-term outcomes leads to large performance gains over conventional approaches that optimize for short-term proxies.
☆ ConsRec: Learning Consensus Behind Interactions for Group Recommendation WWW'2023
Since group activities have become very common in daily life, there is an urgent demand for generating recommendations for a group of users, referred to as group recommendation task. Existing group recommendation methods usually infer groups' preferences via aggregating diverse members' interests. Actually, groups' ultimate choice involves compromises between members, and finally, an agreement can be reached. However, existing individual information aggregation lacks a holistic group-level consideration, failing to capture the consensus information. Besides, their specific aggregation strategies either suffer from high computational costs or become too coarse-grained to make precise predictions. To solve the aforementioned limitations, in this paper, we focus on exploring consensus behind group behavior data. To comprehensively capture the group consensus, we innovatively design three distinct views which provide mutually complementary information to enable multi-view learning, including member-level aggregation, item-level tastes, and group-level inherent preferences. To integrate and balance the multi-view information, an adaptive fusion component is further proposed. As to member-level aggregation, different from existing linear or attentive strategies, we design a novel hypergraph neural network that allows for efficient hypergraph convolutional operations to generate expressive member-level aggregation. We evaluate our ConsRec on two real-world datasets and experimental results show that our model outperforms state-of-the-art methods. An extensive case study also verifies the effectiveness of consensus modeling.
comment: Accepted by WWW'2023
☆ Multi-Task Deep Recommender Systems: A Survey
Multi-task learning (MTL) aims at learning related tasks in a unified model to achieve mutual improvement among tasks considering their shared knowledge. It is an important topic in recommendation due to the demand for multi-task prediction considering performance and efficiency. Although MTL has been well studied and developed, there is still a lack of systematic review in the recommendation community. To fill the gap, we provide a comprehensive review of existing multi-task deep recommender systems (MTDRS) in this survey. To be specific, the problem definition of MTDRS is first given, and it is compared with other related areas. Next, the development of MTDRS is depicted and the taxonomy is introduced from the task relation and methodology aspects. Specifically, the task relation is categorized into parallel, cascaded, and auxiliary with main, while the methodology is grouped into parameter sharing, optimization, and training mechanism. The survey concludes by summarizing the application and public datasets of MTDRS and highlighting the challenges and future directions of the field.
☆ On the Theories Behind Hard Negative Sampling for Recommendation WWW'2023
Negative sampling has been heavily used to train recommender models on large-scale data, wherein sampling hard examples usually not only accelerates the convergence but also improves the model accuracy. Nevertheless, the reasons for the effectiveness of Hard Negative Sampling (HNS) have not been revealed yet. In this work, we fill the research gap by conducting thorough theoretical analyses on HNS. Firstly, we prove that employing HNS on the Bayesian Personalized Ranking (BPR) learner is equivalent to optimizing One-way Partial AUC (OPAUC). Concretely, the BPR equipped with Dynamic Negative Sampling (DNS) is an exact estimator, while with softmax-based sampling is a soft estimator. Secondly, we prove that OPAUC has a stronger connection with Top-K evaluation metrics than AUC and verify it with simulation experiments. These analyses establish the theoretical foundation of HNS in optimizing Top-K recommendation performance for the first time. On these bases, we offer two insightful guidelines for effective usage of HNS: 1) the sampling hardness should be controllable, e.g., via pre-defined hyper-parameters, to adapt to different Top-K metrics and datasets; 2) the smaller the $K$ we emphasize in Top-K evaluation metrics, the harder the negative samples we should draw. Extensive experiments on three real-world benchmarks verify the two guidelines.
comment: Accepted by WWW'2023
☆ Exploration and Regularization of the Latent Action Space in Recommendation
In recommender systems, reinforcement learning solutions have effectively boosted recommendation performance because of their ability to capture long-term user-system interaction. However, the action space of the recommendation policy is a list of items, which could be extremely large with a dynamic candidate item pool. To overcome this challenge, we propose a hyper-actor and critic learning framework where the policy decomposes the item list generation process into a hyper-action inference step and an effect-action selection step. The first step maps the given state space into a vectorized hyper-action space, and the second step selects the item list based on the hyper-action. In order to regulate the discrepancy between the two action spaces, we design an alignment module along with a kernel mapping function for items to ensure inference accuracy and include a supervision module to stabilize the learning process. We build simulated environments on public datasets and empirically show that our framework is superior in recommendation compared to standard RL baselines.
☆ Self-Sampling Training and Evaluation for the Accuracy-Bias Tradeoff in Recommendation DASFAA 2023
Research on debiased recommendation has shown promising results. However, some issues still need to be handled for its application in industrial recommendation. For example, most of the existing methods require some specific data, architectures and training methods. In this paper, we first argue through an online study that arbitrarily removing all the biases in industrial recommendation may not consistently yield a desired performance improvement. For the situation that a randomized dataset is not available, we propose a novel self-sampling training and evaluation (SSTE) framework to achieve the accuracy-bias tradeoff in recommendation, i.e., eliminate the harmful biases and preserve the beneficial ones. Specifically, SSTE uses a self-sampling module to generate some subsets with different degrees of bias from the original training and validation data. A self-training module infers the beneficial biases and learns better tradeoff based on these subsets, and a self-evaluation module aims to use these subsets to construct more plausible references to reflect the optimized model. Finally, we conduct extensive offline experiments on two datasets to verify the effectiveness of our SSTE. Moreover, we deploy our SSTE in homepage recommendation of a famous financial management product called Tencent Licaitong, and find very promising results in an online A/B test.
comment: Accepted by DASFAA 2023 Industry Track
☆ Multi-Task Recommendations with Reinforcement Learning
In recent years, Multi-task Learning (MTL) has yielded immense success in Recommender System (RS) applications. However, current MTL-based recommendation models tend to disregard the session-wise patterns of user-item interactions because they are predominantly constructed based on item-wise datasets. Moreover, balancing multiple objectives has always been a challenge in this field, which is typically avoided via linear estimations in existing works. To address these issues, in this paper, we propose a Reinforcement Learning (RL) enhanced MTL framework, namely RMTL, to combine the losses of different recommendation tasks using dynamic weights. To be specific, the RMTL structure can address the two aforementioned issues by (i) constructing an MTL environment from session-wise interactions and (ii) training multi-task actor-critic network structure, which is compatible with most existing MTL-based recommendation models, and (iii) optimizing and fine-tuning the MTL loss function using the weights generated by critic networks. Experiments on two real-world public datasets demonstrate the effectiveness of RMTL with a higher AUC against state-of-the-art MTL-based recommendation models. Additionally, we evaluate and validate RMTL's compatibility and transferability across various MTL models.
comment: TheWebConf2023
☆ PLACES: Prompting Language Models for Social Conversation Synthesis EACL 2023
Collecting high quality conversational data can be very expensive for most applications and infeasible for others due to privacy, ethical, or similar concerns. A promising direction to tackle this problem is to generate synthetic dialogues by prompting large language models. In this work, we use a small set of expert-written conversations as in-context examples to synthesize a social conversation dataset using prompting. We perform several thorough evaluations of our synthetic conversations compared to human-collected conversations. This includes various dimensions of conversation quality with human evaluation directly on the synthesized conversations, and interactive human evaluation of chatbots fine-tuned on the synthetically generated dataset. We additionally demonstrate that this prompting approach is generalizable to multi-party conversations, providing potential to create new synthetic data for multi-party tasks. Our synthetic multi-party conversations were rated more favorably across all measured dimensions compared to conversation excerpts sampled from a human-collected multi-party dataset.
comment: In EACL 2023. 25 pages, 4 figures, 26 tables. Link to code forthcoming
☆ Disentangled Causal Embedding With Contrastive Learning For Recommender System WWW'23
Recommender systems usually rely on observed user interaction data to build personalized recommendation models, assuming that the observed data reflect user interest. However, user interacting with an item may also due to conformity, the need to follow popular items. Most previous studies neglect user's conformity and entangle interest with it, which may cause the recommender systems fail to provide satisfying results. Therefore, from the cause-effect view, disentangling these interaction causes is a crucial issue. It also contributes to OOD problems, where training and test data are out-of-distribution. Nevertheless, it is quite challenging as we lack the signal to differentiate interest and conformity. The data sparsity of pure cause and the items' long-tail problem hinder disentangled causal embedding. In this paper, we propose DCCL, a framework that adopts contrastive learning to disentangle these two causes by sample augmentation for interest and conformity respectively. Futhermore, DCCL is model-agnostic, which can be easily deployed in any industrial online system. Extensive experiments are conducted over two real-world datasets and DCCL outperforms state-of-the-art baselines on top of various backbone models in various OOD environments. We also demonstrate the performance improvements by online A/B testing on Kuaishou, a billion-user scale short-video recommender system.
comment: Accepted by WWW'23
☆ Towards Lightweight Cross-domain Sequential Recommendation via External Attention-enhanced Graph Convolution Network DASFAA 2023
Cross-domain Sequential Recommendation (CSR) is an emerging yet challenging task that depicts the evolution of behavior patterns for overlapped users by modeling their interactions from multiple domains. Existing studies on CSR mainly focus on using composite or in-depth structures that achieve significant improvement in accuracy but bring a huge burden to the model training. Moreover, to learn the user-specific sequence representations, existing works usually adopt the global relevance weighting strategy (e.g., self-attention mechanism), which has quadratic computational complexity. In this work, we introduce a lightweight external attention-enhanced GCN-based framework to solve the above challenges, namely LEA-GCN. Specifically, by only keeping the neighborhood aggregation component and using the Single-Layer Aggregating Protocol (SLAP), our lightweight GCN encoder performs more efficiently to capture the collaborative filtering signals of the items from both domains. To further alleviate the framework structure and aggregate the user-specific sequential pattern, we devise a novel dual-channel External Attention (EA) component, which calculates the correlation among all items via a lightweight linear structure. Extensive experiments are conducted on two real-world datasets, demonstrating that LEA-GCN requires a smaller volume and less training time without affecting the accuracy compared with several state-of-the-art methods.
comment: 16 pages, 4 figures, conference paper, accepted by DASFAA 2023
♻ ☆ A Counterfactual Collaborative Session-based Recommender System
Most session-based recommender systems (SBRSs) focus on extracting information from the observed items in the current session of a user to predict a next item, ignoring the causes outside the session (called outer-session causes, OSCs) that influence the user's selection of items. However, these causes widely exist in the real world, and few studies have investigated their role in SBRSs. In this work, we analyze the causalities and correlations of the OSCs in SBRSs from the perspective of causal inference. We find that the OSCs are essentially the confounders in SBRSs, which leads to spurious correlations in the data used to train SBRS models. To address this problem, we propose a novel SBRS framework named COCO-SBRS (COunterfactual COllaborative Session-Based Recommender Systems) to learn the causality between OSCs and user-item interactions in SBRSs. COCO-SBRS first adopts a self-supervised approach to pre-train a recommendation model by designing pseudo-labels of causes for each user's selection of the item in data to guide the training process. Next, COCO-SBRS adopts counterfactual inference to recommend items based on the outputs of the pre-trained recommendation model considering the causalities to alleviate the data sparsity problem. As a result, COCO-SBRS can learn the causalities in data, preventing the model from learning spurious correlations. The experimental results of our extensive experiments conducted on three real-world datasets demonstrate the superiority of our proposed framework over ten representative SBRSs.
comment: accepted by the ACM WebConf 2023
♻ ☆ Reinforcing User Retention in a Billion Scale Short Video Recommender System
Recently, short video platforms have achieved rapid user growth by recommending interesting content to users. The objective of the recommendation is to optimize user retention, thereby driving the growth of DAU (Daily Active Users). Retention is a long-term feedback after multiple interactions of users and the system, and it is hard to decompose retention reward to each item or a list of items. Thus traditional point-wise and list-wise models are not able to optimize retention. In this paper, we choose reinforcement learning methods to optimize the retention as they are designed to maximize the long-term performance. We formulate the problem as an infinite-horizon request-based Markov Decision Process, and our objective is to minimize the accumulated time interval of multiple sessions, which is equal to improving the app open frequency and user retention. However, current reinforcement learning algorithms can not be directly applied in this setting due to uncertainty, bias, and long delay time incurred by the properties of user retention. We propose a novel method, dubbed RLUR, to address the aforementioned challenges. Both offline and live experiments show that RLUR can significantly improve user retention. RLUR has been fully launched in Kuaishou app for a long time, and achieves consistent performance improvement on user retention and DAU.
Machine Learning 150
☆ Long Horizon Temperature Scaling
Temperature scaling is a popular technique for tuning the sharpness of a model distribution. It is used extensively for sampling likely generations and calibrating model uncertainty, and even features as a controllable parameter to many large language models in deployment. However, autoregressive models rely on myopic temperature scaling that greedily optimizes the next token. To address this, we propose Long Horizon Temperature Scaling (LHTS), a novel approach for sampling from temperature-scaled joint distributions. LHTS is compatible with all likelihood-based models, and optimizes for the long-horizon likelihood of samples. We derive a temperature-dependent LHTS objective, and show that fine-tuning a model on a range of temperatures produces a single model capable of generation with a controllable long-horizon temperature parameter. We experiment with LHTS on image diffusion models and character/language autoregressive models, demonstrating advantages over myopic temperature scaling in likelihood and sample quality, and showing improvements in accuracy on a multiple choice analogy task by $10\%$.
☆ Temporal Robustness against Data Poisoning
Data poisoning considers cases when an adversary maliciously inserts and removes training data to manipulate the behavior of machine learning algorithms. Traditional threat models of data poisoning center around a single metric, the number of poisoned samples. In consequence, existing defenses are essentially vulnerable in practice when poisoning more samples remains a feasible option for attackers. To address this issue, we leverage timestamps denoting the birth dates of data, which are often available but neglected in the past. Benefiting from these timestamps, we propose a temporal threat model of data poisoning and derive two novel metrics, earliness and duration, which respectively measure how long an attack started in advance and how long an attack lasted. With these metrics, we define the notions of temporal robustness against data poisoning, providing a meaningful sense of protection even with unbounded amounts of poisoned samples. We present a benchmark with an evaluation protocol simulating continuous data collection and periodic deployments of updated models, thus enabling empirical evaluation of temporal robustness. Lastly, we develop and also empirically verify a baseline defense, namely temporal aggregation, offering provable temporal robustness and highlighting the potential of our temporal modeling of data poisoning.
comment: 11 pages, 7 figures
☆ Linear Partial Monitoring for Sequential Decision-Making: Algorithms, Regret Bounds and Applications
Partial monitoring is an expressive framework for sequential decision-making with an abundance of applications, including graph-structured and dueling bandits, dynamic pricing and transductive feedback models. We survey and extend recent results on the linear formulation of partial monitoring that naturally generalizes the standard linear bandit setting. The main result is that a single algorithm, information-directed sampling (IDS), is (nearly) worst-case rate optimal in all finite-action games. We present a simple and unified analysis of stochastic partial monitoring, and further extend the model to the contextual and kernelized setting.
☆ How Reliable is Your Regression Model's Uncertainty Under Real-World Distribution Shifts?
Many important computer vision applications are naturally formulated as regression problems. Within medical imaging, accurate regression models have the potential to automate various tasks, helping to lower costs and improve patient outcomes. Such safety-critical deployment does however require reliable estimation of model uncertainty, also under the wide variety of distribution shifts that might be encountered in practice. Motivated by this, we set out to investigate the reliability of regression uncertainty estimation methods under various real-world distribution shifts. To that end, we propose an extensive benchmark of 8 image-based regression datasets with different types of challenging distribution shifts. We then employ our benchmark to evaluate many of the most common uncertainty estimation methods, as well as two state-of-the-art uncertainty scores from the task of out-of-distribution detection. We find that while methods are well calibrated when there is no distribution shift, they all become highly overconfident on many of the benchmark datasets. This uncovers important limitations of current uncertainty estimation methods, and the proposed benchmark therefore serves as a challenge to the research community. We hope that our benchmark will spur more work on how to develop truly reliable regression uncertainty estimation methods. Code is available at https://github.com/fregu856/regression_uncertainty.
comment: Code is available at https://github.com/fregu856/regression_uncertainty
☆ Breaking the Curse of Multiagents in a Large State Space: RL in Markov Games with Independent Linear Function Approximation
We propose a new model, independent linear Markov game, for multi-agent reinforcement learning with a large state space and a large number of agents. This is a class of Markov games with independent linear function approximation, where each agent has its own function approximation for the state-action value functions that are marginalized by other players' policies. We design new algorithms for learning the Markov coarse correlated equilibria (CCE) and Markov correlated equilibria (CE) with sample complexity bounds that only scale polynomially with each agent's own function class complexity, thus breaking the curse of multiagents. In contrast, existing works for Markov games with function approximation have sample complexity bounds scale with the size of the \emph{joint action space} when specialized to the canonical tabular Markov game setting, which is exponentially large in the number of agents. Our algorithms rely on two key technical innovations: (1) utilizing policy replay to tackle non-stationarity incurred by multiple agents and the use of function approximation; (2) separating learning Markov equilibria and exploration in the Markov games, which allows us to use the full-information no-regret learning oracle instead of the stronger bandit-feedback no-regret learning oracle used in the tabular setting. Furthermore, we propose an iterative-best-response type algorithm that can learn pure Markov Nash equilibria in independent linear Markov potential games. In the tabular case, by adapting the policy replay mechanism for independent linear Markov games, we propose an algorithm with $\widetilde{O}(\epsilon^{-2})$ sample complexity to learn Markov CCE, which improves the state-of-the-art result $\widetilde{O}(\epsilon^{-3})$ in Daskalakis et al. 2022, where $\epsilon$ is the desired accuracy, and also significantly improves other problem parameters.
comment: 51 pages
☆ Hard Prompts Made Easy: Gradient-Based Discrete Optimization for Prompt Tuning and Discovery
The strength of modern generative models lies in their ability to be controlled through text-based prompts. Typical "hard" prompts are made from interpretable words and tokens, and must be hand-crafted by humans. There are also "soft" prompts, which consist of continuous feature vectors. These can be discovered using powerful optimization methods, but they cannot be easily interpreted, re-used across models, or plugged into a text-based interface. We describe an approach to robustly optimize hard text prompts through efficient gradient-based optimization. Our approach automatically generates hard text-based prompts for both text-to-image and text-to-text applications. In the text-to-image setting, the method creates hard prompts for diffusion models, allowing API users to easily generate, discover, and mix and match image concepts without prior knowledge on how to prompt the model. In the text-to-text setting, we show that hard prompts can be automatically discovered that are effective in tuning LMs for classification.
comment: 14 pages, 10 figures, Code is available at \url{https://github.com/YuxinWenRick/hard-prompts-made-easy}
☆ SDYN-GANs: Adversarial Learning Methods for Multistep Generative Models for General Order Stochastic Dynamics
We introduce adversarial learning methods for data-driven generative modeling of the dynamics of $n^{th}$-order stochastic systems. Our approach builds on Generative Adversarial Networks (GANs) with generative model classes based on stable $m$-step stochastic numerical integrators. We introduce different formulations and training methods for learning models of stochastic dynamics based on observation of trajectory samples. We develop approaches using discriminators based on Maximum Mean Discrepancy (MMD), training protocols using conditional and marginal distributions, and methods for learning dynamic responses over different time-scales. We show how our approaches can be used for modeling physical systems to learn force-laws, damping coefficients, and noise-related parameters. The adversarial learning approaches provide methods for obtaining stable generative models for dynamic tasks including long-time prediction and developing simulations for stochastic systems.
comment: 7 figures
☆ Federated Learning with Regularized Client Participation
Federated Learning (FL) is a distributed machine learning approach where multiple clients work together to solve a machine learning task. One of the key challenges in FL is the issue of partial participation, which occurs when a large number of clients are involved in the training process. The traditional method to address this problem is randomly selecting a subset of clients at each communication round. In our research, we propose a new technique and design a novel regularized client participation scheme. Under this scheme, each client joins the learning process every $R$ communication rounds, which we refer to as a meta epoch. We have found that this participation scheme leads to a reduction in the variance caused by client sampling. Combined with the popular FedAvg algorithm (McMahan et al., 2017), it results in superior rates under standard assumptions. For instance, the optimization term in our main convergence bound decreases linearly with the product of the number of communication rounds and the size of the local dataset of each client, and the statistical term scales with step size quadratically instead of linearly (the case for client sampling with replacement), leading to better convergence rate $\mathcal{O}\left(\frac{1}{T^2}\right)$ compared to $\mathcal{O}\left(\frac{1}{T}\right)$, where $T$ is the total number of communication rounds. Furthermore, our results permit arbitrary client availability as long as each client is available for training once per each meta epoch.
comment: 33 pages, 10 figures,1 algorithm, 3 theorems
☆ Riemannian Flow Matching on General Geometries
We propose Riemannian Flow Matching (RFM), a simple yet powerful framework for training continuous normalizing flows on manifolds. Existing methods for generative modeling on manifolds either require expensive simulation, inherently cannot scale to high dimensions, or use approximations to limiting quantities that result in biased objectives. Riemannian Flow Matching bypasses these inconveniences and exhibits multiple benefits over prior approaches: It is completely simulation-free on simple geometries, it does not require divergence computation, and its target vector field is computed in closed form even on general geometries. The key ingredient behind RFM is the construction of a simple kernel function for defining per-sample vector fields, which subsumes existing Euclidean cases. Extending to general geometries, we rely on the use of spectral decompositions to efficiently compute kernel functions. Our method achieves state-of-the-art performance on real-world non-Euclidean datasets, and we showcase, for the first time, tractable training on general geometries, including on triangular meshes and maze-like manifolds with boundaries.
☆ Planted Bipartite Graph Detection
We consider the task of detecting a hidden bipartite subgraph in a given random graph. Specifically, under the null hypothesis, the graph is a realization of an Erd\H{o}s-R\'{e}nyi random graph over $n$ vertices with edge density $q$. Under the alternative, there exists a planted $k_{\mathsf{R}} \times k_{\mathsf{L}}$ bipartite subgraph with edge density $p>q$. We derive asymptotically tight upper and lower bounds for this detection problem in both the dense regime, where $q,p = \Theta\left(1\right)$, and the sparse regime where $q,p = \Theta\left(n^{-\alpha}\right), \alpha \in \left(0,2\right]$. Moreover, we consider a variant of the above problem, where one can only observe a relatively small part of the graph, by using at most $\mathsf{Q}$ edge queries. For this problem, we derive upper and lower bounds in both the dense and sparse regimes.
comment: 30 pages
☆ Toward Face Biometric De-identification using Adversarial Examples AAAI-23
The remarkable success of face recognition (FR) has endangered the privacy of internet users particularly in social media. Recently, researchers turned to use adversarial examples as a countermeasure. In this paper, we assess the effectiveness of using two widely known adversarial methods (BIM and ILLC) for de-identifying personal images. We discovered, unlike previous claims in the literature, that it is not easy to get a high protection success rate (suppressing identification rate) with imperceptible adversarial perturbation to the human visual system. Finally, we found out that the transferability of adversarial examples is highly affected by the training parameters of the network with which they are generated.
comment: Accepted at the AAAI-23 workshop on Artificial Intelligence for Cyber Security (AICS)
☆ Reducing SO(3) Convolutions to SO(2) for Efficient Equivariant GNNs
Graph neural networks that model 3D data, such as point clouds or atoms, are typically desired to be $SO(3)$ equivariant, i.e., equivariant to 3D rotations. Unfortunately equivariant convolutions, which are a fundamental operation for equivariant networks, increase significantly in computational complexity as higher-order tensors are used. In this paper, we address this issue by reducing the $SO(3)$ convolutions or tensor products to mathematically equivalent convolutions in $SO(2)$ . This is accomplished by aligning the node embeddings' primary axis with the edge vectors, which sparsifies the tensor product and reduces the computational complexity from $O(L^6)$ to $O(L^3)$, where $L$ is the degree of the representation. We demonstrate the potential implications of this improvement by proposing the Equivariant Spherical Channel Network (eSCN), a graph neural network utilizing our novel approach to equivariant convolutions, which achieves state-of-the-art results on the large-scale OC-20 dataset.
comment: 17 pages, 10 figures
☆ A Privacy-Preserving Hybrid Federated Learning Framework for Financial Crime Detection
The recent decade witnessed a surge of increase in financial crimes across the public and private sectors, with an average cost of scams of \$102m to financial institutions in 2022. Developing a mechanism for battling financial crimes is an impending task that requires in-depth collaboration from multiple institutions, and yet such collaboration imposed significant technical challenges due to the privacy and security requirements of distributed financial data. For example, consider the Society for Worldwide Interbank Financial Telecommunications (SWIFT) system, which generates 42 million transactions per day across its 11,000 global institutions. Training a detection model of fraudulent transactions requires not only secured SWIFT transactions but also the private account activities of those involved in each transaction from corresponding bank systems. The distributed nature of both samples and features prevents most existing learning systems from being directly adopted to handle the data mining task. In this paper, we collectively address these challenges by proposing a hybrid federated learning system that offers secure and privacy-aware learning and inference for financial crime detection. We conduct extensive empirical studies to evaluate the proposed framework's detection performance and privacy-protection capability, evaluating its robustness against common malicious attacks of collaborative learning. We release our source code at https://github.com/illidanlab/HyFL .
☆ Deep Class-Incremental Learning: A Survey
Deep models, e.g., CNNs and Vision Transformers, have achieved impressive achievements in many vision tasks in the closed world. However, novel classes emerge from time to time in our ever-changing world, requiring a learning system to acquire new knowledge continually. For example, a robot needs to understand new instructions, and an opinion monitoring system should analyze emerging topics every day. Class-Incremental Learning (CIL) enables the learner to incorporate the knowledge of new classes incrementally and build a universal classifier among all seen classes. Correspondingly, when directly training the model with new class instances, a fatal problem occurs -- the model tends to catastrophically forget the characteristics of former ones, and its performance drastically degrades. There have been numerous efforts to tackle catastrophic forgetting in the machine learning community. In this paper, we survey comprehensively recent advances in deep class-incremental learning and summarize these methods from three aspects, i.e., data-centric, model-centric, and algorithm-centric. We also provide a rigorous and unified evaluation of 16 methods in benchmark image classification tasks to find out the characteristics of different algorithms empirically. Furthermore, we notice that the current comparison protocol ignores the influence of memory budget in model storage, which may result in unfair comparison and biased results. Hence, we advocate fair comparison by aligning the memory budget in evaluation, as well as several memory-agnostic performance measures. The source code to reproduce these evaluations is available at https://github.com/zhoudw-zdw/CIL_Survey/
comment: Code is available at https://github.com/zhoudw-zdw/CIL_Survey/
☆ Ethical Considerations for Collecting Human-Centric Image Datasets
Human-centric image datasets are critical to the development of computer vision technologies. However, recent investigations have foregrounded significant ethical issues related to privacy and bias, which have resulted in the complete retraction, or modification, of several prominent datasets. Recent works have tried to reverse this trend, for example, by proposing analytical frameworks for ethically evaluating datasets, the standardization of dataset documentation and curation practices, privacy preservation methodologies, as well as tools for surfacing and mitigating representational biases. Little attention, however, has been paid to the realities of operationalizing ethical data collection. To fill this gap, we present a set of key ethical considerations and practical recommendations for collecting more ethically-minded human-centric image data. Our research directly addresses issues of privacy and bias by contributing to the research community best practices for ethical data collection, covering purpose, privacy and consent, as well as diversity. We motivate each consideration by drawing on lessons from current practices, dataset withdrawals and audits, and analytical ethical frameworks. Our research is intended to augment recent scholarship, representing an important step toward more responsible data curation practices.
☆ Recent advances in the Self-Referencing Embedding Strings (SELFIES) library
String-based molecular representations play a crucial role in cheminformatics applications, and with the growing success of deep learning in chemistry, have been readily adopted into machine learning pipelines. However, traditional string-based representations such as SMILES are often prone to syntactic and semantic errors when produced by generative models. To address these problems, a novel representation, SELF-referencIng Embedded Strings (SELFIES), was proposed that is inherently 100% robust, alongside an accompanying open-source implementation. Since then, we have generalized SELFIES to support a wider range of molecules and semantic constraints and streamlined its underlying grammar. We have implemented this updated representation in subsequent versions of \selfieslib, where we have also made major advances with respect to design, efficiency, and supported features. Hence, we present the current status of \selfieslib (version 2.1.1) in this manuscript.
comment: 11 pages, 2 figures
☆ Can gamification reduce the burden of self-reporting in mHealth applications? Feasibility study using machine learning from smartwatch data to estimate cognitive load
The effectiveness of digital treatments can be measured by requiring patients to self-report their mental and physical state through mobile applications. However, self-reporting can be overwhelming and may cause patients to disengage from the intervention. In order to address this issue, we conduct a feasibility study to explore the impact of gamification on the cognitive burden of self-reporting. Our approach involves the creation of a system to assess cognitive burden through the analysis of photoplethysmography (PPG) signals obtained from a smartwatch. The system is built by collecting PPG data during both cognitively demanding tasks and periods of rest. The obtained data is utilized to train a machine learning model to detect cognitive load (CL). Subsequently, we create two versions of health surveys: a gamified version and a traditional version. Our aim is to estimate the cognitive load experienced by participants while completing these surveys using their mobile devices. We find that CL detector performance can be enhanced via pre-training on stress detection tasks and requires capturing of a minimum 30 seconds of PPG signal to work adequately. For 10 out of 13 participants, a personalized cognitive load detector can achieve an F1 score above 0.7. We find no difference between the gamified and non-gamified mobile surveys in terms of time spent in the state of high cognitive load but participants prefer the gamified version. The average time spent on each question is 5.5 for gamified survey vs 6 seconds for the non-gamified version.
☆ Online Reinforcement Learning with Uncertain Episode Lengths AAAI-2023
Existing episodic reinforcement algorithms assume that the length of an episode is fixed across time and known a priori. In this paper, we consider a general framework of episodic reinforcement learning when the length of each episode is drawn from a distribution. We first establish that this problem is equivalent to online reinforcement learning with general discounting where the learner is trying to optimize the expected discounted sum of rewards over an infinite horizon, but where the discounting function is not necessarily geometric. We show that minimizing regret with this new general discounting is equivalent to minimizing regret with uncertain episode lengths. We then design a reinforcement learning algorithm that minimizes regret with general discounting but acts for the setting with uncertain episode lengths. We instantiate our general bound for different types of discounting, including geometric and polynomial discounting. We also show that we can obtain similar regret bounds even when the uncertainty over the episode lengths is unknown, by estimating the unknown distribution over time. Finally, we compare our learning algorithms with existing value-iteration based episodic RL algorithms in a grid-world environment.
comment: To appear at AAAI-2023
☆ Graph Generation with Destination-Driven Diffusion Mixture
Generation of graphs is a major challenge for real-world tasks that require understanding the complex nature of their non-Euclidean structures. Although diffusion models have achieved notable success in graph generation recently, they are ill-suited for modeling the structural information of graphs since learning to denoise the noisy samples does not explicitly capture the graph topology. To tackle this limitation, we propose a novel generative process that models the topology of graphs by predicting the destination of the process. Specifically, we design the generative process as a mixture of diffusion processes conditioned on the endpoint in the data distribution, which drives the process toward the probable destination. Further, we introduce new training objectives for learning to predict the destination, and discuss the advantages of our generative framework that can explicitly model the graph topology and exploit the inductive bias of the data. Through extensive experimental validation on general graph and 2D/3D molecular graph generation tasks, we show that our method outperforms previous generative models, generating graphs with correct topology with both continuous and discrete features.
☆ Adaptive Aggregation for Safety-Critical Control
Safety has been recognized as the central obstacle to preventing the use of reinforcement learning (RL) for real-world applications. Different methods have been developed to deal with safety concerns in RL. However, learning reliable RL-based solutions usually require a large number of interactions with the environment. Likewise, how to improve the learning efficiency, specifically, how to utilize transfer learning for safe reinforcement learning, has not been well studied. In this work, we propose an adaptive aggregation framework for safety-critical control. Our method comprises two key techniques: 1) we learn to transfer the safety knowledge by aggregating the multiple source tasks and a target task through the attention network; 2) we separate the goal of improving task performance and reducing constraint violations by utilizing a safeguard. Experiment results demonstrate that our algorithm can achieve fewer safety violations while showing better data efficiency compared with several baselines.
☆ Multi-Scale Message Passing Neural PDE Solvers
We propose a novel multi-scale message passing neural network algorithm for learning the solutions of time-dependent PDEs. Our algorithm possesses both temporal and spatial multi-scale resolution features by incorporating multi-scale sequence models and graph gating modules in the encoder and processor, respectively. Benchmark numerical experiments are presented to demonstrate that the proposed algorithm outperforms baselines, particularly on a PDE with a range of spatial and temporal scales.
☆ Local Neural Descriptor Fields: Locally Conditioned Object Representations for Manipulation ICRA 2023
A robot operating in a household environment will see a wide range of unique and unfamiliar objects. While a system could train on many of these, it is infeasible to predict all the objects a robot will see. In this paper, we present a method to generalize object manipulation skills acquired from a limited number of demonstrations, to novel objects from unseen shape categories. Our approach, Local Neural Descriptor Fields (L-NDF), utilizes neural descriptors defined on the local geometry of the object to effectively transfer manipulation demonstrations to novel objects at test time. In doing so, we leverage the local geometry shared between objects to produce a more general manipulation framework. We illustrate the efficacy of our approach in manipulating novel objects in novel poses -- both in simulation and in the real world.
comment: ICRA 2023, Project Page: https://elchun.github.io/lndf/
☆ From Utilitarian to Rawlsian Designs for Algorithmic Fairness
There is a lack of consensus within the literature as to how `fairness' of algorithmic systems can be measured, and different metrics can often be at odds. In this paper, we approach this task by drawing on the ethical frameworks of utilitarianism and John Rawls. Informally, these two theories of distributive justice measure the `good' as either a population's sum of utility, or worst-off outcomes, respectively. We present a parameterized class of objective functions that interpolates between these two (possibly) conflicting notions of the `good'. This class is shown to represent a relaxation of the Rawlsian `veil of ignorance', and its sequence of optimal solutions converges to both a utilitarian and Rawlsian optimum. Several other properties of this class are studied, including: 1) a relationship to regularized optimization, 2) feasibility of consistent estimation, and 3) algorithmic cost. In several real-world datasets, we compute optimal solutions and construct the tradeoff between utilitarian and Rawlsian notions of the `good'. Empirically, we demonstrate that increasing model complexity can manifest strict improvements to both measures of the `good'. This work suggests that the proper degree of `fairness' can be informed by a designer's preferences over the space of induced utilitarian and Rawlsian `good'.
☆ Optimizing Audio Recommendations for the Long-Term: A Reinforcement Learning Perspective
We study the problem of optimizing a recommender system for outcomes that occur over several weeks or months. We begin by drawing on reinforcement learning to formulate a comprehensive model of users' recurring relationships with a recommender system. Measurement, attribution, and coordination challenges complicate algorithm design. We describe careful modeling -- including a new representation of user state and key conditional independence assumptions -- which overcomes these challenges and leads to simple, testable recommender system prototypes. We apply our approach to a podcast recommender system that makes personalized recommendations to hundreds of millions of listeners. A/B tests demonstrate that purposefully optimizing for long-term outcomes leads to large performance gains over conventional approaches that optimize for short-term proxies.
☆ Learning to cooperatively estimate road surface friction
We present a system for estimating the friction of the pavement surface at any curved road section, by arriving at a consensus estimate, based on data from vehicles that have recently passed through that section. This estimate can help following vehicles. To keep costs down, we depend only on standard automotive sensors, such as the IMU, and sensors for the steering angle and wheel speeds. Our system's workflow consists of: (i) processing of measurements from existing vehicular sensors, to implement a virtual sensor that captures the effect of low friction on the vehicle, (ii) transmitting short kinematic summaries from vehicles to a road side unit (RSU), using V2X communication, and (iii) estimating the friction coefficients, by running a machine learning regressor at the RSU, on summaries from individual vehicles, and then combining several such estimates. In designing and implementing our system over a road network, we face two key questions: (i) should each individual road section have a local friction coefficient regressor, or can we use a global regressor that covers all the possible road sections? and (ii) how accurate are the resulting regressor estimates? We test the performance of design variations of our solution, using simulations on the commercial package Dyna4. We consider a single vehicle type with varying levels of tyre wear, and a range of road friction coefficients. We find that: (a) only a marginal loss of accuracy is incurred in using a global regressor as compared to local regressors, (b) the consensus estimate at the RSU has a worst case error of about ten percent, if the combination is based on at least fifty recently passed vehicles, and (c) our regressors have root mean square (RMS) errors that are less than five percent. The RMS error rate of our system is half as that of a commercial friction estimation service.
comment: Submited to the 2023 IEEE CCTA
☆ Convergence rates for momentum stochastic gradient descent with noise of machine learning type
We consider the momentum stochastic gradient descent scheme (MSGD) and its continuous-in-time counterpart in the context of non-convex optimization. We show almost sure exponential convergence of the objective function value for target functions that are Lipschitz continuous and satisfy the Polyak-Lojasiewicz inequality on the relevant domain, and under assumptions on the stochastic noise that are motivated by overparameterized supervised learning applications. Moreover, we optimize the convergence rate over the set of friction parameters and show that the MSGD process almost surely converges.
☆ Two Losses Are Better Than One: Faster Optimization Using a Cheaper Proxy
We present an algorithm for minimizing an objective with hard-to-compute gradients by using a related, easier-to-access function as a proxy. Our algorithm is based on approximate proximal point iterations on the proxy combined with relatively few stochastic gradients from the objective. When the difference between the objective and the proxy is $\delta$-smooth, our algorithm guarantees convergence at a rate matching stochastic gradient descent on a $\delta$-smooth objective, which can lead to substantially better sample efficiency. Our algorithm has many potential applications in machine learning, and provides a principled means of leveraging synthetic data, physics simulators, mixed public and private data, and more.
☆ Towards Robust Inductive Graph Incremental Learning via Experience Replay
Inductive node-wise graph incremental learning is a challenging task due to the dynamic nature of evolving graphs and the dependencies between nodes. In this paper, we propose a novel experience replay framework, called Structure-Evolution-Aware Experience Replay (SEA-ER), that addresses these challenges by leveraging the topological awareness of GNNs and importance reweighting technique. Our framework effectively addresses the data dependency of node prediction problems in evolving graphs, with a theoretical guarantee that supports its effectiveness. Through empirical evaluation, we demonstrate that our proposed framework outperforms the current state-of-the-art GNN experience replay methods on several benchmark datasets, as measured by metrics such as accuracy and forgetting.
☆ Efficient Parametric Approximations of Neural Network Function Space Distance
It is often useful to compactly summarize important properties of model parameters and training data so that they can be used later without storing and/or iterating over the entire dataset. As a specific case, we consider estimating the Function Space Distance (FSD) over a training set, i.e. the average discrepancy between the outputs of two neural networks. We propose a Linearized Activation Function TRick (LAFTR) and derive an efficient approximation to FSD for ReLU neural networks. The key idea is to approximate the architecture as a linear network with stochastic gating. Despite requiring only one parameter per unit of the network, our approach outcompetes other parametric approximations with larger memory requirements. Applied to continual learning, our parametric approximation is competitive with state-of-the-art nonparametric approximations, which require storing many training examples. Furthermore, we show its efficacy in estimating influence functions accurately and detecting mislabeled examples without expensive iterations over the entire dataset.
comment: 21 pages, 5 figures
☆ OPORP: One Permutation + One Random Projection
Consider two $D$-dimensional data vectors (e.g., embeddings): $u, v$. In many embedding-based retrieval (EBR) applications where the vectors are generated from trained models, $D=256\sim 1024$ are common. In this paper, OPORP (one permutation + one random projection) uses a variant of the ``count-sketch'' type of data structures for achieving data reduction/compression. With OPORP, we first apply a permutation on the data vectors. A random vector $r$ is generated i.i.d. with moments: $E(r_i) = 0, E(r_i^2)=1, E(r_i^3) =0, E(r_i^4)=s$. We multiply (as dot product) $r$ with all permuted data vectors. Then we break the $D$ columns into $k$ equal-length bins and aggregate (i.e., sum) the values in each bin to obtain $k$ samples from each data vector. One crucial step is to normalize the $k$ samples to the unit $l_2$ norm. We show that the estimation variance is essentially: $(s-1)A + \frac{D-k}{D-1}\frac{1}{k}\left[ (1-\rho^2)^2 -2A\right]$, where $A\geq 0$ is a function of the data ($u,v$). This formula reveals several key properties: (1) We need $s=1$. (2) The factor $\frac{D-k}{D-1}$ can be highly beneficial in reducing variances. (3) The term $\frac{1}{k}(1-\rho^2)^2$ is actually the asymptotic variance of the classical correlation estimator. We illustrate that by letting the $k$ in OPORP to be $k=1$ and repeat the procedure $m$ times, we exactly recover the work of ``very spars random projections'' (VSRP). This immediately leads to a normalized estimator for VSRP which substantially improves the original estimator of VSRP. In summary, with OPORP, the two key steps: (i) the normalization and (ii) the fixed-length binning scheme, have considerably improved the accuracy in estimating the cosine similarity, which is a routine (and crucial) task in modern embedding-based retrieval (EBR) applications.
☆ Revised Conditional t-SNE: Looking Beyond the Nearest Neighbors
Conditional t-SNE (ct-SNE) is a recent extension to t-SNE that allows removal of known cluster information from the embedding, to obtain a visualization revealing structure beyond label information. This is useful, for example, when one wants to factor out unwanted differences between a set of classes. We show that ct-SNE fails in many realistic settings, namely if the data is well clustered over the labels in the original high-dimensional space. We introduce a revised method by conditioning the high-dimensional similarities instead of the low-dimensional similarities and storing within- and across-label nearest neighbors separately. This also enables the use of recently proposed speedups for t-SNE, improving the scalability. From experiments on synthetic data, we find that our proposed method resolves the considered problems and improves the embedding quality. On real data containing batch effects, the expected improvement is not always there. We argue revised ct-SNE is preferable overall, given its improved scalability. The results also highlight new open questions, such as how to handle distance variations between clusters.
comment: 13 pages, 8 pages supplement, to be published in the Proceedings of the 21st International Symposium on Intelligent Data Analysis (IDA 2023), Springer, 2023
☆ Learning Translation Quality Evaluation on Low Resource Languages from Large Language Models
Learned metrics such as BLEURT have in recent years become widely employed to evaluate the quality of machine translation systems. Training such metrics requires data which can be expensive and difficult to acquire, particularly for lower-resource languages. We show how knowledge can be distilled from Large Language Models (LLMs) to improve upon such learned metrics without requiring human annotators, by creating synthetic datasets which can be mixed into existing datasets, requiring only a corpus of text in the target language. We show that the performance of a BLEURT-like model on lower resource languages can be improved in this way.
☆ Natural Language Processing for Policymaking
Language is the medium for many political activities, from campaigns to news reports. Natural language processing (NLP) uses computational tools to parse text into key information that is needed for policymaking. In this chapter, we introduce common methods of NLP, including text classification, topic modeling, event extraction, and text scaling. We then overview how these methods can be used for policymaking through four major applications including data collection for evidence-based policymaking, interpretation of political decisions, policy communication, and investigation of policy effects. Finally, we highlight some potential limitations and ethical concerns when using NLP for policymaking. This text is from Chapter 7 (pages 141-162) of the Handbook of Computational Social Science for Policy (2023). Open Access on Springer: https://doi.org/10.1007/978-3-031-16624-2
comment: Handbook of Computational Social Science for Policy (2023), Chapter 7 (pages 141-162). Open Access on Springer: https://doi.org/10.1007/978-3-031-16624-2
☆ A Bayesian Optimization approach for calibrating large-scale activity-based transport models
The use of Agent-Based and Activity-Based modeling in transportation is rising due to the capability of addressing complex applications such as disruptive trends (e.g., remote working and automation) or the design and assessment of disaggregated management strategies. Still, the broad adoption of large-scale disaggregate models is not materializing due to the inherently high complexity and computational needs. Activity-based models focused on behavioral theory, for example, may involve hundreds of parameters that need to be calibrated to match the detailed socio-economical characteristics of the population for any case study. This paper tackles this issue by proposing a novel Bayesian Optimization approach incorporating a surrogate model in the form of an improved Random Forest, designed to automate the calibration process of the behavioral parameters. The proposed method is tested on a case study for the city of Tallinn, Estonia, where the model to be calibrated consists of 477 behavioral parameters, using the SimMobility MT software. Satisfactory performance is achieved in the major indicators defined for the calibration process: the error for the overall number of trips is equal to 4% and the average error in the OD matrix is 15.92 vehicles per day.
☆ Med-NCA: Robust and Lightweight Segmentation with Neural Cellular Automata
Access to the proper infrastructure is critical when performing medical image segmentation with Deep Learning. This requirement makes it difficult to run state-of-the-art segmentation models in resource-constrained scenarios like primary care facilities in rural areas and during crises. The recently emerging field of Neural Cellular Automata (NCA) has shown that locally interacting one-cell models can achieve competitive results in tasks such as image generation or segmentations in low-resolution inputs. However, they are constrained by high VRAM requirements and the difficulty of reaching convergence for high-resolution images. To counteract these limitations we propose Med-NCA, an end-to-end NCA training pipeline for high-resolution image segmentation. Our method follows a two-step process. Global knowledge is first communicated between cells across the downscaled image. Following that, patch-based segmentation is performed. Our proposed Med-NCA outperforms the classic UNet by 2% and 3% Dice for hippocampus and prostate segmentation, respectively, while also being 500 times smaller. We also show that Med-NCA is by design invariant with respect to image scale, shape and translation, experiencing only slight performance degradation even with strong shifts; and is robust against MRI acquisition artefacts. Med-NCA enables high-resolution medical image segmentation even on a Raspberry Pi B+, arguably the smallest device able to run PyTorch and that can be powered by a standard power bank.
☆ Robustness Implies Fairness in Casual Algorithmic Recourse
Algorithmic recourse aims to disclose the inner workings of the black-box decision process in situations where decisions have significant consequences, by providing recommendations to empower beneficiaries to achieve a more favorable outcome. To ensure an effective remedy, suggested interventions must not only be low-cost but also robust and fair. This goal is accomplished by providing similar explanations to individuals who are alike. This study explores the concept of individual fairness and adversarial robustness in causal algorithmic recourse and addresses the challenge of achieving both. To resolve the challenges, we propose a new framework for defining adversarially robust recourse. The new setting views the protected feature as a pseudometric and demonstrates that individual fairness is a special case of adversarial robustness. Finally, we introduce the fair robust recourse problem to achieve both desirable properties and show how it can be satisfied both theoretically and empirically.
☆ Diverse Probabilistic Trajectory Forecasting with Admissibility Constraints
Predicting multiple trajectories for road users is important for automated driving systems: ego-vehicle motion planning indeed requires a clear view of the possible motions of the surrounding agents. However, the generative models used for multiple-trajectory forecasting suffer from a lack of diversity in their proposals. To avoid this form of collapse, we propose a novel method for structured prediction of diverse trajectories. To this end, we complement an underlying pretrained generative model with a diversity component, based on a determinantal point process (DPP). We balance and structure this diversity with the inclusion of knowledge-based quality constraints, independent from the underlying generative model. We combine these two novel components with a gating operation, ensuring that the predictions are both diverse and within the drivable area. We demonstrate on the nuScenes driving dataset the relevance of our compound approach, which yields significant improvements in the diversity and the quality of the generated trajectories.
☆ Mind the Gap! Bridging Explainable Artificial Intelligence and Human Understanding with Luhmann's Functional Theory of Communication
Over the past decade explainable artificial intelligence has evolved from a predominantly technical discipline into a field that is deeply intertwined with social sciences. Insights such as human preference for contrastive -- more precisely, counterfactual -- explanations have played a major role in this transition, inspiring and guiding the research in computer science. Other observations, while equally important, have received much less attention. The desire of human explainees to communicate with artificial intelligence explainers through a dialogue-like interaction has been mostly neglected by the community. This poses many challenges for the effectiveness and widespread adoption of such technologies as delivering a single explanation optimised according to some predefined objectives may fail to engender understanding in its recipients and satisfy their unique needs given the diversity of human knowledge and intention. Using insights elaborated by Niklas Luhmann and, more recently, Elena Esposito we apply social systems theory to highlight challenges in explainable artificial intelligence and offer a path forward, striving to reinvigorate the technical research in this direction. This paper aims to demonstrate the potential of systems theoretical approaches to communication in understanding problems and limitations of explainable artificial intelligence.
☆ On the relationship between multivariate splines and infinitely-wide neural networks
We consider multivariate splines and show that they have a random feature expansion as infinitely wide neural networks with one-hidden layer and a homogeneous activation function which is the power of the rectified linear unit. We show that the associated function space is a Sobolev space on a Euclidean ball, with an explicit bound on the norms of derivatives. This link provides a new random feature expansion for multivariate splines that allow efficient algorithms. This random feature expansion is numerically better behaved than usual random Fourier features, both in theory and practice. In particular, in dimension one, we compare the associated leverage scores to compare the two random expansions and show a better scaling for the neural network expansion.
☆ Ensemble Value Functions for Efficient Exploration in Multi-Agent Reinforcement Learning
Cooperative multi-agent reinforcement learning (MARL) requires agents to explore to learn to cooperate. Existing value-based MARL algorithms commonly rely on random exploration, such as $\epsilon$-greedy, which is inefficient in discovering multi-agent cooperation. Additionally, the environment in MARL appears non-stationary to any individual agent due to the simultaneous training of other agents, leading to highly variant and thus unstable optimisation signals. In this work, we propose ensemble value functions for multi-agent exploration (EMAX), a general framework to extend any value-based MARL algorithm. EMAX trains ensembles of value functions for each agent to address the key challenges of exploration and non-stationarity: (1) The uncertainty of value estimates across the ensemble is used in a UCB policy to guide the exploration of agents to parts of the environment which require cooperation. (2) Average value estimates across the ensemble serve as target values. These targets exhibit lower variance compared to commonly applied target networks and we show that they lead to more stable gradients during the optimisation. We instantiate three value-based MARL algorithms with EMAX, independent DQN, VDN and QMIX, and evaluate them in 21 tasks across four environments. Using ensembles of five value functions, EMAX improves sample efficiency and final evaluation returns of these algorithms by 54%, 55%, and 844%, respectively, averaged all 21 tasks.
comment: Preprint. Under review
☆ Uncoupled Learning of Differential Stackelberg Equilibria with Commitments
A natural solution concept for many multiagent settings is the Stackelberg equilibrium, under which a ``leader'' agent selects a strategy that maximizes its own payoff assuming the ``follower'' chooses their best response to this strategy. Recent work has presented asymmetric learning updates that can be shown to converge to the \textit{differential} Stackelberg equilibria of two-player differentiable games. These updates are ``coupled'' in the sense that the leader requires some information about the follower's payoff function. Such coupled learning rules cannot be applied to \textit{ad hoc} interactive learning settings, and can be computationally impractical even in centralized training settings where the follower's payoffs are known. In this work, we present an ``uncoupled'' learning process under which each player's learning update only depends on their observations of the other's behavior. We prove that this process converges to a local Stackelberg equilibrium under similar conditions as previous coupled methods. We conclude with a discussion of the potential applications of our approach to human--AI cooperation and multi-agent reinforcement learning.
☆ Towards Skilled Population Curriculum for Multi-Agent Reinforcement Learning
Recent advances in multi-agent reinforcement learning (MARL) allow agents to coordinate their behaviors in complex environments. However, common MARL algorithms still suffer from scalability and sparse reward issues. One promising approach to resolving them is automatic curriculum learning (ACL). ACL involves a student (curriculum learner) training on tasks of increasing difficulty controlled by a teacher (curriculum generator). Despite its success, ACL's applicability is limited by (1) the lack of a general student framework for dealing with the varying number of agents across tasks and the sparse reward problem, and (2) the non-stationarity of the teacher's task due to ever-changing student strategies. As a remedy for ACL, we introduce a novel automatic curriculum learning framework, Skilled Population Curriculum (SPC), which adapts curriculum learning to multi-agent coordination. Specifically, we endow the student with population-invariant communication and a hierarchical skill set, allowing it to learn cooperation and behavior skills from distinct tasks with varying numbers of agents. In addition, we model the teacher as a contextual bandit conditioned by student policies, enabling a team of agents to change its size while still retaining previously acquired skills. We also analyze the inherent non-stationarity of this multi-agent automatic curriculum teaching problem and provide a corresponding regret bound. Empirical results show that our method improves the performance, scalability and sample efficiency in several MARL environments.
☆ A unified recipe for deriving (time-uniform) PAC-Bayes bounds
We present a unified framework for deriving PAC-Bayesian generalization bounds. Unlike most previous literature on this topic, our bounds are anytime-valid (i.e., time-uniform), meaning that they hold at all stopping times, not only for a fixed sample size. Our approach combines four tools in the following order: (a) nonnegative supermartingales or reverse submartingales, (b) the method of mixtures, (c) the Donsker-Varadhan formula (or other convex duality principles), and (d) Ville's inequality. We derive time-uniform generalizations of well-known classical PAC-Bayes bounds, such as those of Seeger, McAllester, Maurer, and Catoni, in addition to many recent bounds. We also present several novel bounds and, more importantly, general techniques for constructing them. Despite being anytime-valid, our extensions remain as tight as their fixed-time counterparts. Moreover, they enable us to relax traditional assumptions; in particular, we consider nonstationary loss functions and non-i.i.d. data. In sum, we unify the derivation of past bounds and ease the search for future bounds: one may simply check if our supermartingale or submartingale conditions are met and, if so, be guaranteed a (time-uniform) PAC-Bayes bound.
☆ Averaged Method of Multipliers for Bi-Level Optimization without Lower-Level Strong Convexity
Gradient methods have become mainstream techniques for Bi-Level Optimization (BLO) in learning fields. The validity of existing works heavily rely on either a restrictive Lower- Level Strong Convexity (LLSC) condition or on solving a series of approximation subproblems with high accuracy or both. In this work, by averaging the upper and lower level objectives, we propose a single loop Bi-level Averaged Method of Multipliers (sl-BAMM) for BLO that is simple yet efficient for large-scale BLO and gets rid of the limited LLSC restriction. We further provide non-asymptotic convergence analysis of sl-BAMM towards KKT stationary points, and the comparative advantage of our analysis lies in the absence of strong gradient boundedness assumption, which is always required by others. Thus our theory safely captures a wider variety of applications in deep learning, especially where the upper-level objective is quadratic w.r.t. the lower-level variable. Experimental results demonstrate the superiority of our method.
comment: 36 pages
☆ Sparse GEMINI for Joint Discriminative Clustering and Feature Selection
Feature selection in clustering is a hard task which involves simultaneously the discovery of relevant clusters as well as relevant variables with respect to these clusters. While feature selection algorithms are often model-based through optimised model selection or strong assumptions on $p(\pmb{x})$, we introduce a discriminative clustering model trying to maximise a geometry-aware generalisation of the mutual information called GEMINI with a simple $\ell_1$ penalty: the Sparse GEMINI. This algorithm avoids the burden of combinatorial feature subset exploration and is easily scalable to high-dimensional data and large amounts of samples while only designing a clustering model $p_\theta(y|\pmb{x})$. We demonstrate the performances of Sparse GEMINI on synthetic datasets as well as large-scale datasets. Our results show that Sparse GEMINI is a competitive algorithm and has the ability to select relevant subsets of variables with respect to the clustering without using relevance criteria or prior hypotheses.
☆ Learning Discretized Neural Networks under Ricci Flow
In this paper, we consider Discretized Neural Networks (DNNs) consisting of low-precision weights and activations, which suffer from either infinite or zero gradients caused by the non-differentiable discrete function in the training process. In this case, most training-based DNNs use the standard Straight-Through Estimator (STE) to approximate the gradient w.r.t. discrete value. However, the standard STE will cause the gradient mismatch problem, i.e., the approximated gradient direction may deviate from the steepest descent direction. In other words, the gradient mismatch implies the approximated gradient with perturbations. To address this problem, we introduce the duality theory to regard the perturbation of the approximated gradient as the perturbation of the metric in Linearly Nearly Euclidean (LNE) manifolds. Simultaneously, under the Ricci-DeTurck flow, we prove the dynamical stability and convergence of the LNE metric with the $L^2$-norm perturbation, which can provide a theoretical solution for the gradient mismatch problem. In practice, we also present the steepest descent gradient flow for DNNs on LNE manifolds from the viewpoints of the information geometry and mirror descent. The experimental results on various datasets demonstrate that our method achieves better and more stable performance for DNNs than other representative training-based methods.
comment: arXiv admin note: text overlap with arXiv:2111.08410
☆ Phase Transitions in the Detection of Correlated Databases
We study the problem of detecting the correlation between two Gaussian databases $\mathsf{X}\in\mathbb{R}^{n\times d}$ and $\mathsf{Y}^{n\times d}$, each composed of $n$ users with $d$ features. This problem is relevant in the analysis of social media, computational biology, etc. We formulate this as a hypothesis testing problem: under the null hypothesis, these two databases are statistically independent. Under the alternative, however, there exists an unknown permutation $\sigma$ over the set of $n$ users (or, row permutation), such that $\mathsf{X}$ is $\rho$-correlated with $\mathsf{Y}^\sigma$, a permuted version of $\mathsf{Y}$. We determine sharp thresholds at which optimal testing exhibits a phase transition, depending on the asymptotic regime of $n$ and $d$. Specifically, we prove that if $\rho^2d\to0$, as $d\to\infty$, then weak detection (performing slightly better than random guessing) is statistically impossible, irrespectively of the value of $n$. This compliments the performance of a simple test that thresholds the sum all entries of $\mathsf{X}^T\mathsf{Y}$. Furthermore, when $d$ is fixed, we prove that strong detection (vanishing error probability) is impossible for any $\rho<\rho^\star$, where $\rho^\star$ is an explicit function of $d$, while weak detection is again impossible as long as $\rho^2d\to0$. These results close significant gaps in current recent related studies.
comment: 22 pages
☆ Data augmentation for machine learning of chemical process flowsheets SC
Artificial intelligence has great potential for accelerating the design and engineering of chemical processes. Recently, we have shown that transformer-based language models can learn to auto-complete chemical process flowsheets using the SFILES 2.0 string notation. Also, we showed that language translation models can be used to translate Process Flow Diagrams (PFDs) into Process and Instrumentation Diagrams (P&IDs). However, artificial intelligence methods require big data and flowsheet data is currently limited. To mitigate this challenge of limited data, we propose a new data augmentation methodology for flowsheet data that is represented in the SFILES 2.0 notation. We show that the proposed data augmentation improves the performance of artificial intelligence-based process design models. In our case study flowsheet data augmentation improved the prediction uncertainty of the flowsheet autocompletion model by 14.7%. In the future, our flowsheet data augmentation can be used for other machine learning algorithms on chemical process flowsheets that are based on SFILES notation.
comment: Submitted to PROCEEDINGS OF THE 33rd European Symposium on Computer Aided Process Engineering (ESCAPE33), June 18-21, 2023, Athens, Greece
☆ Transfer learning for process design with reinforcement learning SC
Process design is a creative task that is currently performed manually by engineers. Artificial intelligence provides new potential to facilitate process design. Specifically, reinforcement learning (RL) has shown some success in automating process design by integrating data-driven models that learn to build process flowsheets with process simulation in an iterative design process. However, one major challenge in the learning process is that the RL agent demands numerous process simulations in rigorous process simulators, thereby requiring long simulation times and expensive computational power. Therefore, typically short-cut simulation methods are employed to accelerate the learning process. Short-cut methods can, however, lead to inaccurate results. We thus propose to utilize transfer learning for process design with RL in combination with rigorous simulation methods. Transfer learning is an established approach from machine learning that stores knowledge gained while solving one problem and reuses this information on a different target domain. We integrate transfer learning in our RL framework for process design and apply it to an illustrative case study comprising equilibrium reactions, azeotropic separation, and recycles, our method can design economically feasible flowsheets with stable interaction with DWSIM. Our results show that transfer learning enables RL to economically design feasible flowsheets with DWSIM, resulting in a flowsheet with an 8% higher revenue. And the learning time can be reduced by a factor of 2.
comment: Submitted to PROCEEDINGS OF THE 33rd European Symposium on Computer Aided Process Engineering (ESCAPE33), June 18-21, 2023, Athens, Greece
☆ Population-size-Aware Policy Optimization for Mean-Field Games ICLR 2023
In this work, we attempt to bridge the two fields of finite-agent and infinite-agent games, by studying how the optimal policies of agents evolve with the number of agents (population size) in mean-field games, an agent-centric perspective in contrast to the existing works focusing typically on the convergence of the empirical distribution of the population. To this end, the premise is to obtain the optimal policies of a set of finite-agent games with different population sizes. However, either deriving the closed-form solution for each game is theoretically intractable, training a distinct policy for each game is computationally intensive, or directly applying the policy trained in a game to other games is sub-optimal. We address these challenges through the Population-size-Aware Policy Optimization (PAPO). Our contributions are three-fold. First, to efficiently generate efficient policies for games with different population sizes, we propose PAPO, which unifies two natural options (augmentation and hypernetwork) and achieves significantly better performance. PAPO consists of three components: i) the population-size encoding which transforms the original value of population size to an equivalent encoding to avoid training collapse, ii) a hypernetwork to generate a distinct policy for each game conditioned on the population size, and iii) the population size as an additional input to the generated policy. Next, we construct a multi-task-based training procedure to efficiently train the neural networks of PAPO by sampling data from multiple games with different population sizes. Finally, extensive experiments on multiple environments show the significant superiority of PAPO over baselines, and the analysis of the evolution of the generated policies further deepens our understanding of the two fields of finite-agent and infinite-agent games.
comment: Accepted to International Conference on Learning Representations (ICLR 2023)
☆ Machine learning benchmarks for the classification of equivalent circuit models from solid-state electrochemical impedance spectra
Analysis of Electrochemical Impedance Spectroscopy (EIS) data for electrochemical systems often consists of defining an Equivalent Circuit Model (ECM) using expert knowledge and then optimizing the model parameters to deconvolute various resistance, capacitive, inductive, or diffusion responses. For small data sets, this procedure can be conducted manually; however, it is not feasible to manually define a proper ECM for extensive data sets with a wide range of EIS responses. Automatic identification of an ECM would substantially accelerate the analysis of large sets of EIS data. Here, we showcase machine learning methods developed during the BatteryDEV hackathon to classify the ECMs of 9,300 EIS measurements provided by QuantumScape. The best-performing approach is a gradient-boosted tree model utilizing a library to automatically generate features, followed by a random forest model using the raw spectral data. A convolutional neural network using boolean images of Nyquist representations is presented as an alternative, although it achieves a lower accuracy. We publish the data and open source the associated code. The approaches described in this article can serve as benchmarks for further studies. A key remaining challenge is that the labels contain uncertainty and human bias, underlined by the performance of the trained models.
comment: Manuscript: 16 pages, 8 figures; Supplementary Information: 7 pages, 3 figures
☆ A conceptual model for leaving the data-centric approach in machine learning
For a long time, machine learning (ML) has been seen as the abstract problem of learning relationships from data independent of the surrounding settings. This has recently been challenged, and methods have been proposed to include external constraints in the machine learning models. These methods usually come from application-specific fields, such as de-biasing algorithms in the field of fairness in ML or physical constraints in the fields of physics and engineering. In this paper, we present and discuss a conceptual high-level model that unifies these approaches in a common language. We hope that this will enable and foster exchange between the different fields and their different methods for including external constraints into ML models, and thus leaving purely data-centric approaches.
☆ Deep-OSG: A deep learning approach for approximating a family of operators in semigroup to model unknown autonomous systems
This paper proposes a novel deep learning approach for approximating evolution operators and modeling unknown autonomous dynamical systems using time series data collected at varied time lags. It is a sequel to the previous works [T. Qin, K. Wu, and D. Xiu, J. Comput. Phys., 395:620--635, 2019], [K. Wu and D. Xiu, J. Comput. Phys., 408:109307, 2020], and [Z. Chen, V. Churchill, K. Wu, and D. Xiu, J. Comput. Phys., 449:110782, 2022], which focused on learning single evolution operator with a fixed time step. This paper aims to learn a family of evolution operators with variable time steps, which constitute a semigroup for an autonomous system. The semigroup property is very crucial and links the system's evolutionary behaviors across varying time scales, but it was not considered in the previous works. We propose for the first time a framework of embedding the semigroup property into the data-driven learning process, through a novel neural network architecture and new loss functions. The framework is very feasible, can be combined with any suitable neural networks, and is applicable to learning general autonomous ODEs and PDEs. We present the rigorous error estimates and variance analysis to understand the prediction accuracy and robustness of our approach, showing the remarkable advantages of semigroup awareness in our model. Moreover, our approach allows one to arbitrarily choose the time steps for prediction and ensures that the predicted results are well self-matched and consistent. Extensive numerical experiments demonstrate that embedding the semigroup property notably reduces the data dependency of deep learning models and greatly improves the accuracy, robustness, and stability for long-time prediction.
☆ Towards Better Time Series Contrastive Learning: A Dynamic Bad Pair Mining Approach
Not all positive pairs are beneficial to time series contrastive learning. In this paper, we study two types of bad positive pairs that impair the quality of time series representation learned through contrastive learning ($i.e.$, noisy positive pair and faulty positive pair). We show that, with the presence of noisy positive pairs, the model tends to simply learn the pattern of noise (Noisy Alignment). Meanwhile, when faulty positive pairs arise, the model spends considerable efforts aligning non-representative patterns (Faulty Alignment). To address this problem, we propose a Dynamic Bad Pair Mining (DBPM) algorithm, which reliably identifies and suppresses bad positive pairs in time series contrastive learning. DBPM utilizes a memory module to track the training behavior of each positive pair along training process. This allows us to identify potential bad positive pairs at each epoch based on their historical training behaviors. The identified bad pairs are then down-weighted using a transformation module. Our experimental results show that DBPM effectively mitigates the negative impacts of bad pairs, and can be easily used as a plug-in to boost performance of state-of-the-art methods. Codes will be made publicly available.
comment: Preprint. Under review
☆ AMFPMC -- An improved method of detecting multiple types of drug-drug interactions using only known drug-drug interactions
Adverse drug interactions are largely preventable causes of medical accidents, which frequently result in physician and emergency room encounters. The detection of drug interactions in a lab, prior to a drug's use in medical practice, is essential, however it is costly and time-consuming. Machine learning techniques can provide an efficient and accurate means of predicting possible drug-drug interactions and combat the growing problem of adverse drug interactions. Most existing models for predicting interactions rely on the chemical properties of drugs. While such models can be accurate, the required properties are not always available.
☆ An Informative Path Planning Framework for Active Learning in UAV-based Semantic Mapping
Unmanned aerial vehicles (UAVs) are crucial for aerial mapping and general monitoring tasks. Recent progress in deep learning enabled automated semantic segmentation of imagery to facilitate the interpretation of large-scale complex environments. Commonly used supervised deep learning for segmentation relies on large amounts of pixel-wise labelled data, which is tedious and costly to annotate. The domain-specific visual appearance of aerial environments often prevents the usage of models pre-trained on a static dataset. To address this, we propose a novel general planning framework for UAVs to autonomously acquire informative training images for model re-training. We leverage multiple acquisition functions and fuse them into probabilistic terrain maps. Our framework combines the mapped acquisition function information into the UAV's planning objectives. In this way, the UAV adaptively acquires informative aerial images to be manually labelled for model re-training. Experimental results on real-world data and in a photorealistic simulation show that our framework maximises model performance and drastically reduces labelling efforts. Our map-based planners outperform state-of-the-art local planning.
comment: 15 pages, 20 figures
☆ Towards a User Privacy-Aware Mobile Gaming App Installation Prediction Model
Over the past decade, programmatic advertising has received a great deal of attention in the online advertising industry. A real-time bidding (RTB) system is rapidly becoming the most popular method to buy and sell online advertising impressions. Within the RTB system, demand-side platforms (DSP) aim to spend advertisers' campaign budgets efficiently while maximizing profit, seeking impressions that result in high user responses, such as clicks or installs. In the current study, we investigate the process of predicting a mobile gaming app installation from the point of view of a particular DSP, while paying attention to user privacy, and exploring the trade-off between privacy preservation and model performance. There are multiple levels of potential threats to user privacy, depending on the privacy leaks associated with the data-sharing process, such as data transformation or de-anonymization. To address these concerns, privacy-preserving techniques were proposed, such as cryptographic approaches, for training privacy-aware machine-learning models. However, the ability to train a mobile gaming app installation prediction model without using user-level data, can prevent these threats and protect the users' privacy, even though the model's ability to predict may be impaired. Additionally, current laws might force companies to declare that they are collecting data, and might even give the user the option to opt out of such data collection, which might threaten companies' business models in digital advertising, which are dependent on the collection and use of user-level data. We conclude that privacy-aware models might still preserve significant capabilities, enabling companies to make better decisions, dependent on the privacy-efficacy trade-off utility function of each case.
comment: 11 pages, 3 figures
☆ Multi-Task Recommendations with Reinforcement Learning
In recent years, Multi-task Learning (MTL) has yielded immense success in Recommender System (RS) applications. However, current MTL-based recommendation models tend to disregard the session-wise patterns of user-item interactions because they are predominantly constructed based on item-wise datasets. Moreover, balancing multiple objectives has always been a challenge in this field, which is typically avoided via linear estimations in existing works. To address these issues, in this paper, we propose a Reinforcement Learning (RL) enhanced MTL framework, namely RMTL, to combine the losses of different recommendation tasks using dynamic weights. To be specific, the RMTL structure can address the two aforementioned issues by (i) constructing an MTL environment from session-wise interactions and (ii) training multi-task actor-critic network structure, which is compatible with most existing MTL-based recommendation models, and (iii) optimizing and fine-tuning the MTL loss function using the weights generated by critic networks. Experiments on two real-world public datasets demonstrate the effectiveness of RMTL with a higher AUC against state-of-the-art MTL-based recommendation models. Additionally, we evaluate and validate RMTL's compatibility and transferability across various MTL models.
comment: TheWebConf2023
☆ Attacking Cooperative Multi-Agent Reinforcement Learning by Adversarial Minority Influence
Cooperative multi-agent reinforcement learning (c-MARL) offers a general paradigm for a group of agents to achieve a shared goal by taking individual decisions, yet is found to be vulnerable to adversarial attacks. Though harmful, adversarial attacks also play a critical role in evaluating the robustness and finding blind spots of c-MARL algorithms. However, existing attacks are not sufficiently strong and practical, which is mainly due to the ignorance of complex influence between agents and cooperative nature of victims in c-MARL. In this paper, we propose adversarial minority influence (AMI), the first practical attack against c-MARL by introducing an adversarial agent. AMI addresses the aforementioned problems by unilaterally influencing other cooperative victims to a targeted worst-case cooperation. Technically, to maximally deviate victim policy under complex agent-wise influence, our unilateral attack characterize and maximize the influence from adversary to victims. This is done by adapting a unilateral agent-wise relation metric derived from mutual information, which filters out the detrimental influence from victims to adversary. To fool victims into a jointly worst-case failure, our targeted attack influence victims to a long-term, cooperatively worst case by distracting each victim to a specific target. Such target is learned by a reinforcement learning agent in a trial-and-error process. Extensive experiments in simulation environments, including discrete control (SMAC), continuous control (MAMujoco) and real-world robot swarm control demonstrate the superiority of our AMI approach. Our codes are available in https://anonymous.4open.science/r/AMI.
☆ Leveraging Demonstrations to Improve Online Learning: Quality Matters
We investigate the extent to which offline demonstration data can improve online learning. It is natural to expect some improvement, but the question is how, and by how much? We show that the degree of improvement must depend on the quality of the demonstration data. To generate portable insights, we focus on Thompson sampling (TS) applied to a multi-armed bandit as a prototypical online learning algorithm and model. The demonstration data is generated by an expert with a given competence level, a notion we introduce. We propose an informed TS algorithm that utilizes the demonstration data in a coherent way through Bayes' rule and derive a prior-dependent Bayesian regret bound. This offers insight into how pretraining can greatly improve online performance and how the degree of improvement increases with the expert's competence level. We also develop a practical, approximate informed TS algorithm through Bayesian bootstrapping and show substantial empirical regret reduction through experiments.
☆ Federated Variational Inference Methods for Structured Latent Variable Models
Federated learning methods, that is, methods that perform model training using data situated across different sources, whilst simultaneously not having the data leave their original source, are of increasing interest in a number of fields. However, despite this interest, the classes of models for which easily-applicable and sufficiently general approaches are available is limited, excluding many structured probabilistic models. We present a general yet elegant resolution to the aforementioned issue. The approach is based on adopting structured variational inference, an approach widely used in Bayesian machine learning, to the federated setting. Additionally, a communication-efficient variant analogous to the canonical FedAvg algorithm is explored. The effectiveness of the proposed algorithms are demonstrated, and their performance is compared on Bayesian multinomial regression, topic modelling, and mixed model examples.
☆ Mismatched estimation of non-symmetric rank-one matrices corrupted by structured noise
We study the performance of a Bayesian statistician who estimates a rank-one signal corrupted by non-symmetric rotationally invariant noise with a generic distribution of singular values. As the signal-to-noise ratio and the noise structure are unknown, a Gaussian setup is incorrectly assumed. We derive the exact analytic expression for the error of the mismatched Bayes estimator and also provide the analysis of an approximate message passing (AMP) algorithm. The first result exploits the asymptotic behavior of spherical integrals for rectangular matrices and of low-rank matrix perturbations; the second one relies on the design and analysis of an auxiliary AMP. The numerical experiments show that there is a performance gap between the AMP and Bayes estimators, which is due to the incorrect estimation of the signal norm.
☆ Towards Meaningful Anomaly Detection: The Effect of Counterfactual Explanations on the Investigation of Anomalies in Multivariate Time Series
Detecting rare events is essential in various fields, e.g., in cyber security or maintenance. Often, human experts are supported by anomaly detection systems as continuously monitoring the data is an error-prone and tedious task. However, among the anomalies detected may be events that are rare, e.g., a planned shutdown of a machine, but are not the actual event of interest, e.g., breakdowns of a machine. Therefore, human experts are needed to validate whether the detected anomalies are relevant. We propose to support this anomaly investigation by providing explanations of anomaly detection. Related work only focuses on the technical implementation of explainable anomaly detection and neglects the subsequent human anomaly investigation. To address this research gap, we conduct a behavioral experiment using records of taxi rides in New York City as a testbed. Participants are asked to differentiate extreme weather events from other anomalous events such as holidays or sporting events. Our results show that providing counterfactual explanations do improve the investigation of anomalies, indicating potential for explainable anomaly detection in general.
☆ Scalable Gaussian process regression enables accurate prediction of protein and small molecule properties with uncertainty quantitation
Gaussian process (GP) is a Bayesian model which provides several advantages for regression tasks in machine learning such as reliable quantitation of uncertainty and improved interpretability. Their adoption has been precluded by their excessive computational cost and by the difficulty in adapting them for analyzing sequences (e.g. amino acid and nucleotide sequences) and graphs (e.g. ones representing small molecules). In this study, we develop efficient and scalable approaches for fitting GP models as well as fast convolution kernels which scale linearly with graph or sequence size. We implement these improvements by building an open-source Python library called xGPR. We compare the performance of xGPR with the reported performance of various deep learning models on 20 benchmarks, including small molecule, protein sequence and tabular data. We show that xGRP achieves highly competitive performance with much shorter training time. Furthermore, we also develop new kernels for sequence and graph data and show that xGPR generally outperforms convolutional neural networks on predicting key properties of proteins and small molecules. Importantly, xGPR provides uncertainty information not available from typical deep learning models. Additionally, xGPR provides a representation of the input data that can be used for clustering and data visualization. These results demonstrate that xGPR provides a powerful and generic tool that can be broadly useful in protein engineering and drug discovery.
☆ Unsupervised Deep Learning for IoT Time Series
IoT time series analysis has found numerous applications in a wide variety of areas, ranging from health informatics to network security. Nevertheless, the complex spatial temporal dynamics and high dimensionality of IoT time series make the analysis increasingly challenging. In recent years, the powerful feature extraction and representation learning capabilities of deep learning (DL) have provided an effective means for IoT time series analysis. However, few existing surveys on time series have systematically discussed unsupervised DL-based methods. To fill this void, we investigate unsupervised deep learning for IoT time series, i.e., unsupervised anomaly detection and clustering, under a unified framework. We also discuss the application scenarios, public datasets, existing challenges, and future research directions in this area.
comment: 22 pages, 8 figures, has been accepted by IEEE Internet of Things Journal
☆ An End-to-End Two-Phase Deep Learning-Based workflow to Segment Man-made Objects Around Reservoirs
Reservoirs are fundamental infrastructures for the management of water resources. Constructions around them can negatively impact their quality. Such unauthorized constructions can be monitored by land cover mapping (LCM) remote sensing (RS) images. In this paper, we develop a new approach based on DL and image processing techniques for man-made object segmentation around the reservoirs. In order to segment man-made objects around the reservoirs in an end-to-end procedure, segmenting reservoirs and identifying the region of interest (RoI) around them are essential. In the proposed two-phase workflow, the reservoir is initially segmented using a DL model. A post-processing stage is proposed to remove errors such as floating vegetation. Next, the RoI around the reservoir (RoIaR) is identified using the proposed image processing techniques. Finally, the man-made objects in the RoIaR are segmented using a DL architecture. We trained the proposed workflow using collected Google Earth (GE) images of eight reservoirs in Brazil over two different years. The U-Net-based and SegNet-based architectures are trained to segment the reservoirs. To segment man-made objects in the RoIaR, we trained and evaluated four possible architectures, U-Net, FPN, LinkNet, and PSPNet. Although the collected data has a high diversity (for example, they belong to different states, seasons, resolutions, etc.), we achieved good performances in both phases. Furthermore, applying the proposed post-processing to the output of reservoir segmentation improves the precision in all studied reservoirs except two cases. We validated the prepared workflow with a reservoir dataset outside the training reservoirs. The results show high generalization ability of the prepared workflow.
comment: 21 pages, 13 figures
☆ Utility-based Perturbed Gradient Descent: An Optimizer for Continual Learning
Modern representation learning methods may fail to adapt quickly under non-stationarity since they suffer from the problem of catastrophic forgetting and decaying plasticity. Such problems prevent learners from fast adaptation to changes since they result in increasing numbers of saturated features and forgetting useful features when presented with new experiences. Hence, these methods are rendered ineffective for continual learning. This paper proposes Utility-based Perturbed Gradient Descent (UPGD), an online representation-learning algorithm well-suited for continual learning agents with no knowledge about task boundaries. UPGD protects useful weights or features from forgetting and perturbs less useful ones based on their utilities. Our empirical results show that UPGD alleviates catastrophic forgetting and decaying plasticity, enabling modern representation learning methods to work in the continual learning setting.
☆ IB-UQ: Information bottleneck based uncertainty quantification for neural function regression and neural operator learning
In this paper, a novel framework is established for uncertainty quantification via information bottleneck (IB-UQ) for scientific machine learning tasks, including deep neural network (DNN) regression and neural operator learning (DeepONet). Specifically, we first employ the General Incompressible-Flow Networks (GIN) model to learn a "wide" distribution fromnoisy observation data. Then, following the information bottleneck objective, we learn a stochastic map from input to some latent representation that can be used to predict the output. A tractable variational bound on the IB objective is constructed with a normalizing flow reparameterization. Hence, we can optimize the objective using the stochastic gradient descent method. IB-UQ can provide both mean and variance in the label prediction by explicitly modeling the representation variables. Compared to most DNN regression methods and the deterministic DeepONet, the proposed model can be trained on noisy data and provide accurate predictions with reliable uncertainty estimates on unseen noisy data. We demonstrate the capability of the proposed IB-UQ framework via several representative examples, including discontinuous function regression, real-world dataset regression and learning nonlinear operators for diffusion-reaction partial differential equation.
comment: 22 pages, 36figures
☆ Learning to Count Isomorphisms with Graph Neural Networks
Subgraph isomorphism counting is an important problem on graphs, as many graph-based tasks exploit recurring subgraph patterns. Classical methods usually boil down to a backtracking framework that needs to navigate a huge search space with prohibitive computational costs. Some recent studies resort to graph neural networks (GNNs) to learn a low-dimensional representation for both the query and input graphs, in order to predict the number of subgraph isomorphisms on the input graph. However, typical GNNs employ a node-centric message passing scheme that receives and aggregates messages on nodes, which is inadequate in complex structure matching for isomorphism counting. Moreover, on an input graph, the space of possible query graphs is enormous, and different parts of the input graph will be triggered to match different queries. Thus, expecting a fixed representation of the input graph to match diversely structured query graphs is unrealistic. In this paper, we propose a novel GNN called Count-GNN for subgraph isomorphism counting, to deal with the above challenges. At the edge level, given that an edge is an atomic unit of encoding graph structures, we propose an edge-centric message passing scheme, where messages on edges are propagated and aggregated based on the edge adjacency to preserve fine-grained structural information. At the graph level, we modulate the input graph representation conditioned on the query, so that the input graph can be adapted to each query individually to improve their matching. Finally, we conduct extensive experiments on a number of benchmark datasets to demonstrate the superior performance of Count-GNN.
☆ Delving Deep into Simplicity Bias for Long-Tailed Image Recognition
Simplicity Bias (SB) is a phenomenon that deep neural networks tend to rely favorably on simpler predictive patterns but ignore some complex features when applied to supervised discriminative tasks. In this work, we investigate SB in long-tailed image recognition and find the tail classes suffer more severely from SB, which harms the generalization performance of such underrepresented classes. We empirically report that self-supervised learning (SSL) can mitigate SB and perform in complementary to the supervised counterpart by enriching the features extracted from tail samples and consequently taking better advantage of such rare samples. However, standard SSL methods are designed without explicitly considering the inherent data distribution in terms of classes and may not be optimal for long-tailed distributed data. To address this limitation, we propose a novel SSL method tailored to imbalanced data. It leverages SSL by triple diverse levels, i.e., holistic-, partial-, and augmented-level, to enhance the learning of predictive complex patterns, which provides the potential to overcome the severe SB on tail data. Both quantitative and qualitative experimental results on five long-tailed benchmark datasets show our method can effectively mitigate SB and significantly outperform the competing state-of-the-arts.
☆ Membership Inference Attacks against Diffusion Models
Diffusion models have attracted attention in recent years as innovative generative models. In this paper, we investigate whether a diffusion model is resistant to a membership inference attack, which evaluates the privacy leakage of a machine learning model. We primarily discuss the diffusion model from the standpoints of comparison with a generative adversarial network (GAN) as conventional models and hyperparameters unique to the diffusion model, i.e., time steps, sampling steps, and sampling variances. We conduct extensive experiments with DDIM as a diffusion model and DCGAN as a GAN on the CelebA and CIFAR-10 datasets in both white-box and black-box settings and then confirm if the diffusion model is comparably resistant to a membership inference attack as GAN. Next, we demonstrate that the impact of time steps is significant and intermediate steps in a noise schedule are the most vulnerable to the attack. We also found two key insights through further analysis. First, we identify that DDIM is vulnerable to the attack for small sample sizes instead of achieving a lower FID. Second, sampling steps in hyperparameters are important for resistance to the attack, whereas the impact of sampling variances is quite limited.
☆ Climate Intervention Analysis using AI Model Guided by Statistical Physics Principles
The availability of training data remains a significant obstacle for the implementation of machine learning in scientific applications. In particular, estimating how a system might respond to external forcings or perturbations requires specialized labeled data or targeted simulations, which may be computationally intensive to generate at scale. In this study, we propose a novel solution to this challenge by utilizing a principle from statistical physics known as the Fluctuation-Dissipation Theorem (FDT) to discover knowledge using an AI model that can rapidly produce scenarios for different external forcings. By leveraging FDT, we are able to extract information encoded in a large dataset produced by Earth System Models, which includes 8250 years of internal climate fluctuations, to estimate the climate system's response to forcings. Our model, AiBEDO, is capable of capturing the complex, multi-timescale effects of radiation perturbations on global and regional surface climate, allowing for a substantial acceleration of the exploration of the impacts of spatially-heterogenous climate forcers. To demonstrate the utility of AiBEDO, we use the example of a climate intervention technique called Marine Cloud Brightening, with the ultimate goal of optimizing the spatial pattern of cloud brightening to achieve regional climate targets and prevent known climate tipping points. While we showcase the effectiveness of our approach in the context of climate science, it is generally applicable to other scientific disciplines that are limited by the extensive computational demands of domain simulation models. Source code of AiBEDO framework is made available at https://github.com/kramea/kdd_aibedo. A sample dataset is made available at https://doi.org/10.5281/zenodo.7597027. Additional data available upon request.
☆ DivBO: Diversity-aware CASH for Ensemble Learning NeurIPS 2022
The Combined Algorithm Selection and Hyperparameters optimization (CASH) problem is one of the fundamental problems in Automated Machine Learning (AutoML). Motivated by the success of ensemble learning, recent AutoML systems build post-hoc ensembles to output the final predictions instead of using the best single learner. However, while most CASH methods focus on searching for a single learner with the best performance, they neglect the diversity among base learners (i.e., they may suggest similar configurations to previously evaluated ones), which is also a crucial consideration when building an ensemble. To tackle this issue and further enhance the ensemble performance, we propose DivBO, a diversity-aware framework to inject explicit search of diversity into the CASH problems. In the framework, we propose to use a diversity surrogate to predict the pair-wise diversity of two unseen configurations. Furthermore, we introduce a temporary pool and a weighted acquisition function to guide the search of both performance and diversity based on Bayesian optimization. Empirical results on 15 public datasets show that DivBO achieves the best average ranks (1.82 and 1.73) on both validation and test errors among 10 compared methods, including post-hoc designs in recent AutoML systems and state-of-the-art baselines for ensemble learning on CASH problems.
comment: with appendix, accepted by NeurIPS 2022
☆ CDANs: Temporal Causal Discovery from Autocorrelated and Non-Stationary Time Series Data
This study presents a novel constraint-based causal discovery approach for autocorrelated and non-stationary time series data (CDANs). Our proposed method addresses several limitations of existing causal discovery methods for autocorrelated and non-stationary time series data, such as high dimensionality, the inability to identify lagged causal relationships, and the overlook of changing modules. Our approach identifies both lagged and instantaneous/contemporaneous causal relationships along with changing modules that vary over time. The method optimizes the conditioning sets in a constraint-based search by considering lagged parents instead of conditioning on the entire past that addresses high dimensionality. The changing modules are detected by considering both contemporaneous and lagged parents. The approach first detects the lagged adjacencies, then identifies the changing modules and contemporaneous adjacencies, and finally determines the causal direction. We extensively evaluated the proposed method using synthetic datasets and a real-world clinical dataset and compared its performance with several baseline approaches. The results demonstrate the effectiveness of the proposed method in detecting causal relationships and changing modules in autocorrelated and non-stationary time series data.
☆ Quantum Recurrent Neural Networks for Sequential Learning
Quantum neural network (QNN) is one of the promising directions where the near-term noisy intermediate-scale quantum (NISQ) devices could find advantageous applications against classical resources. Recurrent neural networks are the most fundamental networks for sequential learning, but up to now there is still a lack of canonical model of quantum recurrent neural network (QRNN), which certainly restricts the research in the field of quantum deep learning. In the present work, we propose a new kind of QRNN which would be a good candidate as the canonical QRNN model, where, the quantum recurrent blocks (QRBs) are constructed in the hardware-efficient way, and the QRNN is built by stacking the QRBs in a staggered way that can greatly reduce the algorithm's requirement with regard to the coherent time of quantum devices. That is, our QRNN is much more accessible on NISQ devices. Furthermore, the performance of the present QRNN model is verified concretely using three different kinds of classical sequential data, i.e., meteorological indicators, stock price, and text categorization. The numerical experiments show that our QRNN achieves much better performance in prediction (classification) accuracy against the classical RNN and state-of-the-art QNN models for sequential learning, and can predict the changing details of temporal sequence data. The practical circuit structure and superior performance indicate that the present QRNN is a promising learning model to find quantum advantageous applications in the near term.
☆ Continual Learning of Language Models ICLR 2023
Language models (LMs) have been instrumental for the rapid advance of natural language processing. This paper studies continual learning of LMs, in particular, continual domain-adaptive pre-training (or continual DAP-training). Existing research has shown that further pre-training an LM using a domain corpus to adapt the LM to the domain can improve the end-task performance in the domain. This paper proposes a novel method to continually DAP-train an LM with a sequence of unlabeled domain corpora to adapt the LM to these domains to improve their end-task performances. The key novelty of our method is a soft-masking mechanism that directly controls the update to the LM. A novel proxy is also proposed to preserve the general knowledge in the original LM. Additionally, it contrasts the representations of the previously learned domain knowledge (including the general knowledge in the pre-trained LM) and the knowledge from the current full network to achieve knowledge integration. The method not only overcomes catastrophic forgetting, but also achieves knowledge transfer to improve end-task performances. Empirical evaluation demonstrates the effectiveness of the proposed method.
comment: ICLR 2023
☆ Decentralized Inexact Proximal Gradient Method With Network-Independent Stepsizes for Convex Composite Optimization
This paper considers decentralized convex composite optimization over undirected and connected networks, where the local loss function contains both smooth and nonsmooth terms. For this problem, a novel CTA (Combine-Then-Adapt)-based decentralized algorithm is proposed under uncoordinated network-independent constant stepsizes. Particularly, the proposed algorithm only needs to approximately solve a sequence of proximal mappings, which benefits the decentralized composite optimization where the proximal mappings of the nonsmooth loss functions may not have analytic solutions. For the general convex case, we prove the O(1/k) convergence rate of the proposed algorithm, which can be improved to o(1/k) if the proximal mappings are solved exactly. Moreover, with metric subregularity, we establish the linear convergence rate. Finally, the numerical experiments demonstrate the efficiency of the algorithm.
☆ Exact Inference in High-order Structured Prediction
In this paper, we study the problem of inference in high-order structured prediction tasks. In the context of Markov random fields, the goal of a high-order inference task is to maximize a score function on the space of labels, and the score function can be decomposed into sum of unary and high-order potentials. We apply a generative model approach to study the problem of high-order inference, and provide a two-stage convex optimization algorithm for exact label recovery. We also provide a new class of hypergraph structural properties related to hyperedge expansion that drives the success in general high-order inference problems. Finally, we connect the performance of our algorithm and the hyperedge expansion property using a novel hypergraph Cheeger-type inequality.
☆ Linear optimal partial transport embedding
Optimal transport (OT) has gained popularity due to its various applications in fields such as machine learning, statistics, and signal processing. However, the balanced mass requirement limits its performance in practical problems. To address these limitations, variants of the OT problem, including unbalanced OT, Optimal partial transport (OPT), and Hellinger Kantorovich (HK), have been proposed. In this paper, we propose the Linear optimal partial transport (LOPT) embedding, which extends the (local) linearization technique on OT and HK to the OPT problem. The proposed embedding allows for faster computation of OPT distance between pairs of positive measures. Besides our theoretical contributions, we demonstrate the LOPT embedding technique in point-cloud interpolation and PCA analysis.
☆ Heterophily-Aware Graph Attention Network
Graph Neural Networks (GNNs) have shown remarkable success in graph representation learning. Unfortunately, current weight assignment schemes in standard GNNs, such as the calculation based on node degrees or pair-wise representations, can hardly be effective in processing the networks with heterophily, in which the connected nodes usually possess different labels or features. Existing heterophilic GNNs tend to ignore the modeling of heterophily of each edge, which is also a vital part in tackling the heterophily problem. In this paper, we firstly propose a heterophily-aware attention scheme and reveal the benefits of modeling the edge heterophily, i.e., if a GNN assigns different weights to edges according to different heterophilic types, it can learn effective local attention patterns, which enable nodes to acquire appropriate information from distinct neighbors. Then, we propose a novel Heterophily-Aware Graph Attention Network (HA-GAT) by fully exploring and utilizing the local distribution as the underlying heterophily, to handle the networks with different homophily ratios. To demonstrate the effectiveness of the proposed HA-GAT, we analyze the proposed heterophily-aware attention scheme and local distribution exploration, by seeking for an interpretation from their mechanism. Extensive results demonstrate that our HA-GAT achieves state-of-the-art performances on eight datasets with different homophily ratios in both the supervised and semi-supervised node classification tasks.
☆ Automatic Sleep Stage Classification with Cross-modal Self-supervised Features from Deep Brain Signals
The detection of human sleep stages is widely used in the diagnosis and intervention of neurological and psychiatric diseases. Some patients with deep brain stimulator implanted could have their neural activities recorded from the deep brain. Sleep stage classification based on deep brain recording has great potential to provide more precise treatment for patients. The accuracy and generalizability of existing sleep stage classifiers based on local field potentials are still limited. We proposed an applicable cross-modal transfer learning method for sleep stage classification with implanted devices. This end-to-end deep learning model contained cross-modal self-supervised feature representation, self-attention, and classification framework. We tested the model with deep brain recording data from 12 patients with Parkinson's disease. The best total accuracy reached 83.2% for sleep stage classification. Results showed speech self-supervised features catch the conversion pattern of sleep stages effectively. We provide a new method on transfer learning from acoustic signals to local field potentials. This method supports an effective solution for the insufficient scale of clinical data. This sleep stage classification model could be adapted to chronic and continuous monitor sleep for Parkinson's patients in daily life, and potentially utilized for more precise treatment in deep brain-machine interfaces, such as closed-loop deep brain stimulation.
comment: 4 pages, 5 figures, 11th International IEEE EMBS Conference on Neural Engineering (NER)
☆ Efficient XAI Techniques: A Taxonomic Survey
Recently, there has been a growing demand for the deployment of Explainable Artificial Intelligence (XAI) algorithms in real-world applications. However, traditional XAI methods typically suffer from a high computational complexity problem, which discourages the deployment of real-time systems to meet the time-demanding requirements of real-world scenarios. Although many approaches have been proposed to improve the efficiency of XAI methods, a comprehensive understanding of the achievements and challenges is still needed. To this end, in this paper we provide a review of efficient XAI. Specifically, we categorize existing techniques of XAI acceleration into efficient non-amortized and efficient amortized methods. The efficient non-amortized methods focus on data-centric or model-centric acceleration upon each individual instance. In contrast, amortized methods focus on learning a unified distribution of model explanations, following the predictive, generative, or reinforcement frameworks, to rapidly derive multiple model explanations. We also analyze the limitations of an efficient XAI pipeline from the perspectives of the training phase, the deployment phase, and the use scenarios. Finally, we summarize the challenges of deploying XAI acceleration methods to real-world scenarios, overcoming the trade-off between faithfulness and efficiency, and the selection of different acceleration methods.
comment: 14 pages, 3 figures
☆ Undersampling and Cumulative Class Re-decision Methods to Improve Detection of Agitation in People with Dementia
Agitation is one of the most prevalent symptoms in people with dementia (PwD) that can place themselves and the caregiver's safety at risk. Developing objective agitation detection approaches is important to support health and safety of PwD living in a residential setting. In a previous study, we collected multimodal wearable sensor data from 17 participants for 600 days and developed machine learning models for predicting agitation in one-minute windows. However, there are significant limitations in the dataset, such as imbalance problem and potential imprecise labels as the occurrence of agitation is much rarer in comparison to the normal behaviours. In this paper, we first implement different undersampling methods to eliminate the imbalance problem, and come to the conclusion that only 20% of normal behaviour data are adequate to train a competitive agitation detection model. Then, we design a weighted undersampling method to evaluate the manual labeling mechanism given the ambiguous time interval (ATI) assumption. After that, the postprocessing method of cumulative class re-decision (CCR) is proposed based on the historical sequential information and continuity characteristic of agitation, improving the decision-making performance for the potential application of agitation detection system. The results show that a combination of undersampling and CCR improves best F1-score by 26.6% and other metrics to varying degrees with less training time and data used, and inspires a way to find the potential range of optimal threshold reference for clinical purpose.
comment: 19 pages, 8 figures
☆ Towards Lightweight Cross-domain Sequential Recommendation via External Attention-enhanced Graph Convolution Network DASFAA 2023
Cross-domain Sequential Recommendation (CSR) is an emerging yet challenging task that depicts the evolution of behavior patterns for overlapped users by modeling their interactions from multiple domains. Existing studies on CSR mainly focus on using composite or in-depth structures that achieve significant improvement in accuracy but bring a huge burden to the model training. Moreover, to learn the user-specific sequence representations, existing works usually adopt the global relevance weighting strategy (e.g., self-attention mechanism), which has quadratic computational complexity. In this work, we introduce a lightweight external attention-enhanced GCN-based framework to solve the above challenges, namely LEA-GCN. Specifically, by only keeping the neighborhood aggregation component and using the Single-Layer Aggregating Protocol (SLAP), our lightweight GCN encoder performs more efficiently to capture the collaborative filtering signals of the items from both domains. To further alleviate the framework structure and aggregate the user-specific sequential pattern, we devise a novel dual-channel External Attention (EA) component, which calculates the correlation among all items via a lightweight linear structure. Extensive experiments are conducted on two real-world datasets, demonstrating that LEA-GCN requires a smaller volume and less training time without affecting the accuracy compared with several state-of-the-art methods.
comment: 16 pages, 4 figures, conference paper, accepted by DASFAA 2023
☆ LUT-NN: Towards Unified Neural Network Inference by Table Lookup
DNN inference requires huge effort of system development and resource cost. This drives us to propose LUT-NN, the first trial towards empowering deep neural network (DNN) inference by table lookup, to eliminate the diverse computation kernels as well as save running cost. Based on the feature similarity of each layer, LUT-NN can learn the typical features, named centroids, of each layer from the training data, precompute them with model weights, and save the results in tables. For future input, the results of the closest centroids with the input features can be directly read from the table, as the approximation of layer output. We propose the novel centroid learning technique for DNN, which enables centroid learning through backpropagation, and adapts three levels of approximation to minimize the model loss. By this technique, LUT-NN achieves comparable accuracy (<5% difference) with original models on real complex dataset, including CIFAR, ImageNet, and GLUE. LUT-NN simplifies the computing operators to only two: closest centroid search and table lookup. We implement them for Intel and ARM CPUs. The model size is reduced by up to 3.5x for CNN models and 7x for BERT. Latency-wise, the real speedup of LUT-NN is up to 7x for BERT and 2x for ResNet, much lower than theoretical results because of the current unfriendly hardware design for table lookup. We expect firstclass table lookup support in the future to unleash the potential of LUT-NN.
☆ Near-Minimax-Optimal Risk-Sensitive Reinforcement Learning with CVaR
In this paper, we study risk-sensitive Reinforcement Learning (RL), focusing on the objective of Conditional Value at Risk (CVaR) with risk tolerance $\tau$. Starting with multi-arm bandits (MABs), we show the minimax CVaR regret rate is $\Omega(\sqrt{\tau^{-1}AK})$, where $A$ is the number of actions and $K$ is the number of episodes, and that it is achieved by an Upper Confidence Bound algorithm with a novel Bernstein bonus. For online RL in tabular Markov Decision Processes (MDPs), we show a minimax regret lower bound of $\Omega(\sqrt{\tau^{-1}SAK})$ (with normalized cumulative rewards), where $S$ is the number of states, and we propose a novel bonus-driven Value Iteration procedure. We show that our algorithm achieves the optimal regret of $\widetilde O(\sqrt{\tau^{-1}SAK})$ under a continuity assumption and in general attains a near-optimal regret of $\widetilde O(\tau^{-1}\sqrt{SAK})$, which is minimax-optimal for constant $\tau$. This improves on the best available bounds. By discretizing rewards appropriately, our algorithms are computationally efficient.
☆ On the Ideal Number of Groups for Isometric Gradient Propagation
Recently, various normalization layers have been proposed to stabilize the training of deep neural networks. Among them, group normalization is a generalization of layer normalization and instance normalization by allowing a degree of freedom in the number of groups it uses. However, to determine the optimal number of groups, trial-and-error-based hyperparameter tuning is required, and such experiments are time-consuming. In this study, we discuss a reasonable method for setting the number of groups. First, we find that the number of groups influences the gradient behavior of the group normalization layer. Based on this observation, we derive the ideal number of groups, which calibrates the gradient scale to facilitate gradient descent optimization. Our proposed number of groups is theoretically grounded, architecture-aware, and can provide a proper value in a layer-wise manner for all layers. The proposed method exhibited improved performance over existing methods in numerous neural network architectures, tasks, and datasets.
comment: 10 pages, 2 figures
♻ ☆ Incentive-aware Contextual Pricing with Non-parametric Market Noise
We consider a dynamic pricing problem for repeated contextual second-price auctions with multiple strategic buyers who aim to maximize their long-term time discounted utility. The seller has limited information on buyers' overall demand curves which depends on a non-parametric market-noise distribution, and buyers may potentially submit corrupted bids (relative to true valuations) to manipulate the seller's pricing policy for more favorable reserve prices in the future. We focus on designing the seller's learning policy to set contextual reserve prices where the seller's goal is to minimize regret compared to the revenue of a benchmark clairvoyant policy that has full information of buyers' demand. We propose a policy with a phased-structure that incorporates randomized "isolation" periods, during which a buyer is randomly chosen to solely participate in the auction. We show that this design allows the seller to control the number of periods in which buyers significantly corrupt their bids. We then prove that our policy enjoys a $T$-period regret of $\widetilde{\mathcal{O}}(\sqrt{T})$ facing strategic buyers. Finally, we conduct numerical simulations to compare our proposed algorithm to standard pricing policies. Our numerical results show that our algorithm outperforms these policies under various buyer bidding behavior.
♻ ☆ Iterated Block Particle Filter for High-dimensional Parameter Learning: Beating the Curse of Dimensionality
Parameter learning for high-dimensional, partially observed, and nonlinear stochastic processes is a methodological challenge. Spatiotemporal disease transmission systems provide examples of such processes giving rise to open inference problems. We propose the iterated block particle filter (IBPF) algorithm for learning high-dimensional parameters over graphical state space models with general state spaces, measures, transition densities and graph structure. Theoretical performance guarantees are obtained on beating the curse of dimensionality (COD), algorithm convergence, and likelihood maximization. Experiments on a highly nonlinear and non-Gaussian spatiotemporal model for measles transmission reveal that the iterated ensemble Kalman filter algorithm (Li et al. (2020)) is ineffective and the iterated filtering algorithm (Ionides et al. (2015)) suffers from the COD, while our IBPF algorithm beats COD consistently across various experiments with different metrics.
♻ ☆ Representation Theory for Geometric Quantum Machine Learning
Recent advances in classical machine learning have shown that creating models with inductive biases encoding the symmetries of a problem can greatly improve performance. Importation of these ideas, combined with an existing rich body of work at the nexus of quantum theory and symmetry, has given rise to the field of Geometric Quantum Machine Learning (GQML). Following the success of its classical counterpart, it is reasonable to expect that GQML will play a crucial role in developing problem-specific and quantum-aware models capable of achieving a computational advantage. Despite the simplicity of the main idea of GQML -- create architectures respecting the symmetries of the data -- its practical implementation requires a significant amount of knowledge of group representation theory. We present an introduction to representation theory tools from the optics of quantum learning, driven by key examples involving discrete and continuous groups. These examples are sewn together by an exposition outlining the formal capture of GQML symmetries via "label invariance under the action of a group representation", a brief (but rigorous) tour through finite and compact Lie group representation theory, a reexamination of ubiquitous tools like Haar integration and twirling, and an overview of some successful strategies for detecting symmetries.
comment: 43 pages, 10 figures. Updated to add relevant references
♻ ☆ Supervised Metric Learning to Rank for Retrieval via Contextual Similarity Optimization
There is extensive interest in metric learning methods for image retrieval. Many metric learning loss functions focus on learning a correct ranking of training samples, but strongly overfit semantically inconsistent labels and require a large amount of data. To address these shortcomings, we propose a new metric learning method, called contextual loss, which optimizes contextual similarity in addition to cosine similarity. Our contextual loss implicitly enforces semantic consistency among neighbors while converging to the correct ranking. We empirically show that the proposed loss is more robust to label noise, and is less prone to overfitting even when a large portion of train data is withheld. Extensive experiments demonstrate that our method achieves a new state-of-the-art across four image retrieval benchmarks and multiple different evaluation settings. Code is available at: https://github.com/Chris210634/metric-learning-using-contextual-similarity
♻ ☆ Understanding Why Generalized Reweighting Does Not Improve Over ERM ICLR 2023
Empirical risk minimization (ERM) is known in practice to be non-robust to distributional shift where the training and the test distributions are different. A suite of approaches, such as importance weighting, and variants of distributionally robust optimization (DRO), have been proposed to solve this problem. But a line of recent work has empirically shown that these approaches do not significantly improve over ERM in real applications with distribution shift. The goal of this work is to obtain a comprehensive theoretical understanding of this intriguing phenomenon. We first posit the class of Generalized Reweighting (GRW) algorithms, as a broad category of approaches that iteratively update model parameters based on iterative reweighting of the training samples. We show that when overparameterized models are trained under GRW, the resulting models are close to that obtained by ERM. We also show that adding small regularization which does not greatly affect the empirical training accuracy does not help. Together, our results show that a broad category of what we term GRW approaches are not able to achieve distributionally robust generalization. Our work thus has the following sobering takeaway: to make progress towards distributionally robust generalization, we either have to develop non-GRW approaches, or perhaps devise novel classification/regression loss functions that are adapted to the class of GRW approaches.
comment: ICLR 2023. 40 pages, 3 figures
♻ ☆ Subtyping patients with chronic disease using longitudinal BMI patterns
Obesity is a major health problem, increasing the risk of various major chronic diseases, such as diabetes, cancer, and stroke. While the role of obesity identified by cross-sectional BMI recordings has been heavily studied, the role of BMI trajectories is much less explored. In this study, we use a machine-learning approach to subtype individuals' risk of developing 18 major chronic diseases by using their BMI trajectories extracted from a large and geographically diverse EHR dataset capturing the health status of around two million individuals for a period of six years. We define nine new interpretable and evidence-based variables based on the BMI trajectories to cluster the patients into subgroups using the k-means clustering method. We thoroughly review each cluster's characteristics in terms of demographic, socioeconomic, and physiological measurement variables to specify the distinct properties of the patients in the clusters. In our experiments, the direct relationship of obesity with diabetes, hypertension, Alzheimer's, and dementia has been re-established and distinct clusters with specific characteristics for several of the chronic diseases have been found to be conforming or complementary to the existing body of knowledge.
comment: https://ieeexplore.ieee.org/document/10018864
♻ ☆ Mitigating Algorithmic Bias with Limited Annotations
Existing work on fairness modeling commonly assumes that sensitive attributes for all instances are fully available, which may not be true in many real-world applications due to the high cost of acquiring sensitive information. When sensitive attributes are not disclosed or available, it is needed to manually annotate a small part of the training data to mitigate bias. However, the skewed distribution across different sensitive groups preserves the skewness of the original dataset in the annotated subset, which leads to non-optimal bias mitigation. To tackle this challenge, we propose Active Penalization Of Discrimination (APOD), an interactive framework to guide the limited annotations towards maximally eliminating the effect of algorithmic bias. The proposed APOD integrates discrimination penalization with active instance selection to efficiently utilize the limited annotation budget, and it is theoretically proved to be capable of bounding the algorithmic bias. According to the evaluation on five benchmark datasets, APOD outperforms the state-of-the-arts baseline methods under the limited annotation budget, and shows comparable performance to fully annotated bias mitigation, which demonstrates that APOD could benefit real-world applications when sensitive information is limited.
♻ ☆ Provably Bounding Neural Network Preimages
Most work on the formal verification of neural networks has focused on bounding forward images of neural networks, i.e., the set of outputs of a neural network that correspond to a given set of inputs (for example, bounded perturbations of a nominal input). However, many use cases of neural network verification require solving the inverse problem, i.e, over-approximating the set of inputs that lead to certain outputs. In this work, we present the first efficient bound propagation algorithm, INVPROP, for verifying properties over the preimage of a linearly constrained output set of a neural network, which can be combined with branch-and-bound to achieve completeness. Our efficient algorithm allows multiple passes of intermediate bound refinements, which are crucial for tight inverse verification because the bounds of an intermediate layer depend on relaxations both before and after this layer. We demonstrate our algorithm on applications related to quantifying safe control regions for a dynamical system and detecting out-of-distribution inputs to a neural network. Our results show that in certain settings, we can find over-approximations that are over 2500 times tighter than prior work while being 2.5 times faster on the same hardware.
♻ ☆ Elucidating Robust Learning with Uncertainty-Aware Corruption Pattern Estimation
Robust learning methods aim to learn a clean target distribution from noisy and corrupted training data where a specific corruption pattern is often assumed a priori. Our proposed method can not only successfully learn the clean target distribution from a dirty dataset but also can estimate the underlying noise pattern. To this end, we leverage a mixture-of-experts model that can distinguish two different types of predictive uncertainty, aleatoric and epistemic uncertainty. We show that the ability to estimate the uncertainty plays a significant role in elucidating the corruption patterns as these two objectives are tightly intertwined. We also present a novel validation scheme for evaluating the performance of the corruption pattern estimation. Our proposed method is extensively assessed in terms of both robustness and corruption pattern estimation through a number of domains, including computer vision and natural language processing.
♻ ☆ Online Bayesian Meta-Learning for Cognitive Tracking Radar
A key component of cognitive radar is the ability to generalize, or achieve consistent performance across a range of sensing environments, since aspects of the physical scene may vary over time. This presents a challenge for learning-based waveform selection approaches, since transmission policies which are effective in one scene may be highly suboptimal in another. We address this problem by strategically biasing a learning algorithm by exploiting high-level structure across tracking instances, referred to as meta-learning. In this work, we develop an online meta-learning approach for waveform-agile tracking. This approach uses information gained from previous target tracks to speed up and enhance learning in new tracking instances. This results in sample-efficient learning across a class of finite state target channels by exploiting inherent similarity across tracking scenes, attributed to common physical elements such as target type or clutter statistics. We formulate the online waveform selection problem within the framework of Bayesian learning, and provide prior-dependent performance bounds for the meta-learning problem using Probability Approximately Correct (PAC)-Bayes theory. We present a computationally feasible meta-posterior sampling algorithm and study the performance in a simulation study consisting of diverse scenes. Finally, we examine the potential performance benefits and practical challenges associated with online meta-learning for waveform-agile tracking.
comment: 14 pages, 5 figures
♻ ☆ Meta-Learning Biologically Plausible Plasticity Rules with Random Feedback Pathways
Backpropagation is widely used to train artificial neural networks, but its relationship to synaptic plasticity in the brain is unknown. Some biological models of backpropagation rely on feedback projections that are symmetric with feedforward connections, but experiments do not corroborate the existence of such symmetric backward connectivity. Random feedback alignment offers an alternative model in which errors are propagated backward through fixed, random backward connections. This approach successfully trains shallow models, but learns slowly and does not perform well with deeper models or online learning. In this study, we develop a meta-learning approach to discover interpretable, biologically plausible plasticity rules that improve online learning performance with fixed random feedback connections. The resulting plasticity rules show improved online training of deep models in the low data regime. Our results highlight the potential of meta-learning to discover effective, interpretable learning rules satisfying biological constraints.
♻ ☆ Take One Gram of Neural Features, Get Enhanced Group Robustness ECCV 2022
Predictive performance of machine learning models trained with empirical risk minimization (ERM) can degrade considerably under distribution shifts. The presence of spurious correlations in training datasets leads ERM-trained models to display high loss when evaluated on minority groups not presenting such correlations. Extensive attempts have been made to develop methods improving worst-group robustness. However, they require group information for each training input or at least, a validation set with group labels to tune their hyperparameters, which may be expensive to get or unknown a priori. In this paper, we address the challenge of improving group robustness without group annotation during training or validation. To this end, we propose to partition the training dataset into groups based on Gram matrices of features extracted by an ``identification'' model and to apply robust optimization based on these pseudo-groups. In the realistic context where no group labels are available, our experiments show that our approach not only improves group robustness over ERM but also outperforms all recent baselines
comment: Long version (Previous version: OOD-CV Workshop @ ECCV 2022)
♻ ☆ Performative Reinforcement Learning
We introduce the framework of performative reinforcement learning where the policy chosen by the learner affects the underlying reward and transition dynamics of the environment. Following the recent literature on performative prediction~\cite{Perdomo et. al., 2020}, we introduce the concept of performatively stable policy. We then consider a regularized version of the reinforcement learning problem and show that repeatedly optimizing this objective converges to a performatively stable policy under reasonable assumptions on the transition dynamics. Our proof utilizes the dual perspective of the reinforcement learning problem and may be of independent interest in analyzing the convergence of other algorithms with decision-dependent environments. We then extend our results for the setting where the learner just performs gradient ascent steps instead of fully optimizing the objective, and for the setting where the learner has access to a finite number of trajectories from the changed environment. For both settings, we leverage the dual formulation of performative reinforcement learning and establish convergence to a stable solution. Finally, through extensive experiments on a grid-world environment, we demonstrate the dependence of convergence on various parameters e.g. regularization, smoothness, and the number of samples.
♻ ☆ MA2QL: A Minimalist Approach to Fully Decentralized Multi-Agent Reinforcement Learning
Decentralized learning has shown great promise for cooperative multi-agent reinforcement learning (MARL). However, non-stationarity remains a significant challenge in fully decentralized learning. In the paper, we tackle the non-stationarity problem in the simplest and fundamental way and propose multi-agent alternate Q-learning (MA2QL), where agents take turns updating their Q-functions by Q-learning. MA2QL is a minimalist approach to fully decentralized cooperative MARL but is theoretically grounded. We prove that when each agent guarantees $\varepsilon$-convergence at each turn, their joint policy converges to a Nash equilibrium. In practice, MA2QL only requires minimal changes to independent Q-learning (IQL). We empirically evaluate MA2QL on a variety of cooperative multi-agent tasks. Results show MA2QL consistently outperforms IQL, which verifies the effectiveness of MA2QL, despite such minimal changes.
comment: 18 pages
♻ ☆ Contrastive Learning for Unsupervised Domain Adaptation of Time Series ICLR 2023
Unsupervised domain adaptation (UDA) aims at learning a machine learning model using a labeled source domain that performs well on a similar yet different, unlabeled target domain. UDA is important in many applications such as medicine, where it is used to adapt risk scores across different patient cohorts. In this paper, we develop a novel framework for UDA of time series data, called CLUDA. Specifically, we propose a contrastive learning framework to learn contextual representations in multivariate time series, so that these preserve label information for the prediction task. In our framework, we further capture the variation in the contextual representations between source and target domain via a custom nearest-neighbor contrastive learning. To the best of our knowledge, ours is the first framework to learn domain-invariant, contextual representation for UDA of time series data. We evaluate our framework using a wide range of time series datasets to demonstrate its effectiveness and show that it achieves state-of-the-art performance for time series UDA.
comment: Published as a conference paper at ICLR 2023
♻ ☆ Black Box Optimization Using QUBO and the Cross Entropy Method
Black-box optimization (BBO) can be used to optimize functions whose analytic form is unknown. A common approach to realising BBO is to learn a surrogate model which approximates the target black-box function which can then be solved via white-box optimization methods. In this paper, we present our approach BOX-QUBO, where the surrogate model is a QUBO matrix. However, unlike in previous state-of-the-art approaches, this matrix is not trained entirely by regression, but mostly by classification between 'good' and 'bad' solutions. This better accounts for the low capacity of the QUBO matrix, resulting in significantly better solutions overall. We tested our approach against the state-of-the-art on four domains and in all of them BOX-QUBO showed better results. A second contribution of this paper is the idea to also solve white-box problems, i.e. problems which could be directly formulated as QUBO, by means of black-box optimization in order to reduce the size of the QUBOs to the information-theoretic minimum. Experiments show that this significantly improves the results for MAX-k-SAT.
♻ ☆ Gradient-based Bi-level Optimization for Deep Learning: A Survey
Bi-level optimization, especially the gradient-based category, has been widely used in the deep learning community including hyperparameter optimization and meta knowledge extraction. Bi-level optimization embeds one problem within another and the gradient-based category solves the outer level task by computing the hypergradient, which is much more efficient than classical methods such as the evolutionary algorithm. In this survey, we first give a formal definition of the gradient-based bi-level optimization. Secondly, we illustrate how to formulate a research problem as a bi-level optimization problem, which is of great practical use for beginners. More specifically, there are two formulations: the single-task formulation to optimize hyperparameters such as regularization parameters and the distilled data, and the multi-task formulation to extract meta knowledge such as the model initialization. With a bi-level formulation, we then discuss four bi-level optimization solvers to update the outer variable including explicit gradient update, proxy update, implicit function update, and closed-form update. Last but not least, we conclude the survey by pointing out the great potential of gradient-based bi-level optimization on science problems (AI4Science).
comment: AI4Science; Bi-level Optimization; Hyperparameter Optimization; Meta Learning; Implicit Function
♻ ☆ Self-learning Machines based on Hamiltonian Echo Backpropagation
A physical self-learning machine can be defined as a nonlinear dynamical system that can be trained on data (similar to artificial neural networks), but where the update of the internal degrees of freedom that serve as learnable parameters happens autonomously. In this way, neither external processing and feedback nor knowledge of (and control of) these internal degrees of freedom is required. We introduce a general scheme for self-learning in any time-reversible Hamiltonian system. We illustrate the training of such a self-learning machine numerically for the case of coupled nonlinear wave fields.
♻ ☆ Feature Necessity & Relevancy in ML Classifier Explanations
Given a machine learning (ML) model and a prediction, explanations can be defined as sets of features which are sufficient for the prediction. In some applications, and besides asking for an explanation, it is also critical to understand whether sensitive features can occur in some explanation, or whether a non-interesting feature must occur in all explanations. This paper starts by relating such queries respectively with the problems of relevancy and necessity in logic-based abduction. The paper then proves membership and hardness results for several families of ML classifiers. Afterwards the paper proposes concrete algorithms for two classes of classifiers. The experimental results confirm the scalability of the proposed algorithms.
♻ ☆ A Non-monotonic Self-terminating Language Model ICLR 2023
Recent large-scale neural autoregressive sequence models have shown impressive performances on a variety of natural language generation tasks. However, their generated sequences often exhibit degenerate properties such as non-termination, undesirable repetition, and premature termination, when generated with decoding algorithms such as greedy search, beam search, top-$k$ sampling, and nucleus sampling. In this paper, we focus on the problem of non-terminating sequences resulting from an incomplete decoding algorithm. We first define an incomplete probable decoding algorithm which includes greedy search, top-$k$ sampling, and nucleus sampling, beyond the incomplete decoding algorithm originally put forward by Welleck et al. (2020). We then propose a non-monotonic self-terminating language model, which significantly relaxes the constraint of monotonically increasing termination probability in the originally proposed self-terminating language model by Welleck et al. (2020), to address the issue of non-terminating sequences when using incomplete probable decoding algorithms. We prove that our proposed model prevents non-terminating sequences when using not only incomplete probable decoding algorithms but also beam search. We empirically validate our model on sequence completion tasks with various architectures.
comment: Published as a conference paper at ICLR 2023
♻ ☆ N-Gram Nearest Neighbor Machine Translation
Nearest neighbor machine translation augments the Autoregressive Translation~(AT) with $k$-nearest-neighbor retrieval, by comparing the similarity between the token-level context representations of the target tokens in the query and the datastore. However, the token-level representation may introduce noise when translating ambiguous words, or fail to provide accurate retrieval results when the representation generated by the model contains indistinguishable context information, e.g., Non-Autoregressive Translation~(NAT) models. In this paper, we propose a novel $n$-gram nearest neighbor retrieval method that is model agnostic and applicable to both AT and NAT models. Specifically, we concatenate the adjacent $n$-gram hidden representations as the key, while the tuple of corresponding target tokens is the value. In inference, we propose tailored decoding algorithms for AT and NAT models respectively. We demonstrate that the proposed method consistently outperforms the token-level method on both AT and NAT models as well on general as on domain adaptation translation tasks. On domain adaptation, the proposed method brings $1.03$ and $2.76$ improvements regarding the average BLEU score on AT and NAT models respectively.
♻ ☆ On the symmetries in the dynamics of wide two-layer neural networks
We consider the idealized setting of gradient flow on the population risk for infinitely wide two-layer ReLU neural networks (without bias), and study the effect of symmetries on the learned parameters and predictors. We first describe a general class of symmetries which, when satisfied by the target function $f^*$ and the input distribution, are preserved by the dynamics. We then study more specific cases. When $f^*$ is odd, we show that the dynamics of the predictor reduces to that of a (non-linearly parameterized) linear predictor, and its exponential convergence can be guaranteed. When $f^*$ has a low-dimensional structure, we prove that the gradient flow PDE reduces to a lower-dimensional PDE. Furthermore, we present informal and numerical arguments that suggest that the input neurons align with the lower-dimensional structure of the problem.
♻ ☆ Symmetric Pruning in Quantum Neural Networks ICLR
Many fundamental properties of a quantum system are captured by its Hamiltonian and ground state. Despite the significance of ground states preparation (GSP), this task is classically intractable for large-scale Hamiltonians. Quantum neural networks (QNNs), which exert the power of modern quantum machines, have emerged as a leading protocol to conquer this issue. As such, how to enhance the performance of QNNs becomes a crucial topic in GSP. Empirical evidence showed that QNNs with handcraft symmetric ansatzes generally experience better trainability than those with asymmetric ansatzes, while theoretical explanations have not been explored. To fill this knowledge gap, here we propose the effective quantum neural tangent kernel (EQNTK) and connect this concept with over-parameterization theory to quantify the convergence of QNNs towards the global optima. We uncover that the advance of symmetric ansatzes attributes to their large EQNTK value with low effective dimension, which requests few parameters and quantum circuit depth to reach the over-parameterization regime permitting a benign loss landscape and fast convergence. Guided by EQNTK, we further devise a symmetric pruning (SP) scheme to automatically tailor a symmetric ansatz from an over-parameterized and asymmetric one to greatly improve the performance of QNNs when the explicit symmetry information of Hamiltonian is unavailable. Extensive numerical simulations are conducted to validate the analytical results of EQNTK and the effectiveness of SP.
comment: Accepted to International Conference on Learning Representations (ICLR) 2023
♻ ☆ Universal Soldier: Using Universal Adversarial Perturbations for Detecting Backdoor Attacks
Deep learning models achieve excellent performance in numerous machine learning tasks. Yet, they suffer from security-related issues such as adversarial examples and poisoning (backdoor) attacks. A deep learning model may be poisoned by training with backdoored data or by modifying inner network parameters. Then, a backdoored model performs as expected when receiving a clean input, but it misclassifies when receiving a backdoored input stamped with a pre-designed pattern called "trigger". Unfortunately, it is difficult to distinguish between clean and backdoored models without prior knowledge of the trigger. This paper proposes a backdoor detection method by utilizing a special type of adversarial attack, universal adversarial perturbation (UAP), and its similarities with a backdoor trigger. We observe an intuitive phenomenon: UAPs generated from backdoored models need fewer perturbations to mislead the model than UAPs from clean models. UAPs of backdoored models tend to exploit the shortcut from all classes to the target class, built by the backdoor trigger. We propose a novel method called Universal Soldier for Backdoor detection (USB) and reverse engineering potential backdoor triggers via UAPs. Experiments on 345 models trained on several datasets show that USB effectively detects the injected backdoor and provides comparable or better results than state-of-the-art methods.
♻ ☆ Backpropagation on Dynamical Networks
Dynamical networks are versatile models that can describe a variety of behaviours such as synchronisation and feedback. However, applying these models in real world contexts is difficult as prior information pertaining to the connectivity structure or local dynamics is often unknown and must be inferred from time series observations of network states. Additionally, the influence of coupling interactions between nodes further complicates the isolation of local node dynamics. Given the architectural similarities between dynamical networks and recurrent neural networks (RNN), we propose a network inference method based on the backpropagation through time (BPTT) algorithm commonly used to train recurrent neural networks. This method aims to simultaneously infer both the connectivity structure and local node dynamics purely from observation of node states. An approximation of local node dynamics is first constructed using a neural network. This is alternated with an adapted BPTT algorithm to regress corresponding network weights by minimising prediction errors of the dynamical network based on the previously constructed local models until convergence is achieved. This method was found to be succesful in identifying the connectivity structure for coupled networks of Lorenz, Chua and FitzHugh-Nagumo oscillators. Freerun prediction performance with the resulting local models and weights was found to be comparable to the true system with noisy initial conditions. The method is also extended to non-conventional network couplings such as asymmetric negative coupling.
♻ ☆ Diffusion Denoising for Low-Dose-CT Model
Low-dose Computed Tomography (LDCT) reconstruction is an important task in medical image analysis. Recent years have seen many deep learning based methods, proved to be effective in this area. However, these methods mostly follow a supervised architecture, which needs paired CT image of full dose and quarter dose, and the solution is highly dependent on specific measurements. In this work, we introduce Denoising Diffusion LDCT Model, dubbed as DDLM, generating noise-free CT image using conditioned sampling. DDLM uses pretrained model, and need no training nor tuning process, thus our proposal is in unsupervised manner. Experiments on LDCT images have shown comparable performance of DDLM using less inference time, surpassing other state-of-the-art methods, proving both accurate and efficient. Implementation code will be set to public soon.
comment: The method and experiment of this paper has some error, and we need to revise it
♻ ☆ Equivariant Representation Learning via Class-Pose Decomposition
We introduce a general method for learning representations that are equivariant to symmetries of data. Our central idea is to decompose the latent space into an invariant factor and the symmetry group itself. The components semantically correspond to intrinsic data classes and poses respectively. The learner is trained on a loss encouraging equivariance based on supervision from relative symmetry information. The approach is motivated by theoretical results from group theory and guarantees representations that are lossless, interpretable and disentangled. We provide an empirical investigation via experiments involving datasets with a variety of symmetries. Results show that our representations capture the geometry of data and outperform other equivariant representation learning frameworks.
comment: 12 pages
♻ ☆ Novel Fundus Image Preprocessing for Retcam Images to Improve Deep Learning Classification of Retinopathy of Prematurity
Retinopathy of Prematurity (ROP) is a potentially blinding eye disorder because of damage to the eye's retina which can affect babies born prematurely. Screening of ROP is essential for early detection and treatment. This is a laborious and manual process which requires trained physician performing dilated ophthalmological examination which can be subjective resulting in lower diagnosis success for clinically significant disease. Automated diagnostic methods can assist ophthalmologists increase diagnosis accuracy using deep learning. Several research groups have highlighted various approaches. This paper proposes the use of new novel fundus preprocessing methods using pretrained transfer learning frameworks to create hybrid models to give higher diagnosis accuracy. The evaluations show that these novel methods in comparison to traditional imaging processing contribute to higher accuracy in classifying Plus disease, Stages of ROP and Zones. We achieve accuracy of 97.65% for Plus disease, 89.44% for Stage, 90.24% for Zones with limited training dataset.
comment: 10 pages, 4 figures, 7 tables. arXiv admin note: text overlap with arXiv:1904.08796 by other authors
♻ ☆ FedVeca: Federated Vectorized Averaging on Non-IID Data with Adaptive Bi-directional Global Objective
Federated Learning (FL) is a distributed machine learning framework to alleviate the data silos, where decentralized clients collaboratively learn a global model without sharing their private data. However, the clients' Non-Independent and Identically Distributed (Non-IID) data negatively affect the trained model, and clients with different numbers of local updates may cause significant gaps to the local gradients in each communication round. In this paper, we propose a Federated Vectorized Averaging (FedVeca) method to address the above problem on Non-IID data. Specifically, we set a novel objective for the global model which is related to the local gradients. The local gradient is defined as a bi-directional vector with step size and direction, where the step size is the number of local updates and the direction is divided into positive and negative according to our definition. In FedVeca, the direction is influenced by the step size, thus we average the bi-directional vectors to reduce the effect of different step sizes. Then, we theoretically analyze the relationship between the step sizes and the global objective, and obtain upper bounds on the step sizes per communication round. Based on the upper bounds, we design an algorithm for the server and the client to adaptively adjusts the step sizes that make the objective close to the optimum. Finally, we conduct experiments on different datasets, models and scenarios by building a prototype system, and the experimental results demonstrate the effectiveness and efficiency of the FedVeca method.
♻ ☆ LExecutor: Learning-Guided Execution
Executing code is essential for various program analysis tasks, e.g., to detect bugs that manifest through exceptions or to obtain execution traces for further dynamic analysis. However, executing an arbitrary piece of code is often difficult in practice, e.g., because of missing variable definitions, missing user inputs, and missing third-party dependencies. This paper presents LExecutor, a learning-guided approach for executing arbitrary code snippets in an underconstrained way. The key idea is to let a neural model predict missing values that otherwise would cause the program to get stuck, and to inject these values into the execution. For example, LExecutor injects likely values for otherwise undefined variables and likely return values of calls to otherwise missing functions. We evaluate the approach on Python code from popular open-source projects and on code snippets extracted from Stack Overflow. The neural model predicts realistic values with an accuracy between 80.1% and 94.2%, allowing LExecutor to closely mimic real executions. As a result, the approach successfully executes significantly more code than any available technique, such as simply executing the code as-is. For example, executing the open-source code snippets as-is covers only 4.1% of all lines, because the code crashes early on, whereas LExecutor achieves a coverage of 50.1%.
♻ ☆ A Fuzzy-set-based Joint Distribution Adaptation Method for Regression and its Application to Online Damage Quantification for Structural Digital Twin
Online damage quantification suffers from insufficient labeled data that weakens its accuracy. In this context, adopting the domain adaptation on historical labeled data from similar structures/damages or simulated digital twin data to assist the current diagnosis task would be beneficial. However, most domain adaptation methods are designed for classification and cannot efficiently address damage quantification, a regression problem with continuous real-valued labels. This study first proposes a novel domain adaptation method, the Online Fuzzy-set-based Joint Distribution Adaptation for Regression, to address this challenge. By converting the continuous real-valued labels to fuzzy class labels via fuzzy sets, the marginal and conditional distribution discrepancy are simultaneously measured to achieve the domain adaptation for the damage quantification task. Thanks to the superiority of the proposed method, a state-of-the-art online damage quantification framework based on domain adaptation is presented. Finally, the framework has been comprehensively demonstrated with a damaged helicopter panel, in which three types of damage domain adaptations (across different damage locations, across different damage types, and from simulation to experiment) are all conducted, proving the accuracy of damage quantification can be significantly improved in a realistic environment. It is expected that the proposed approach to be applied to the fleet-level digital twin considering the individual differences.
comment: 29 pages, 10 figures
♻ ☆ Distribution estimation and change-point estimation for time series via DNN-based GANs
The generative adversarial networks (GANs) have recently been applied to estimating the distribution of independent and identically distributed data, and have attracted a lot of research attention. In this paper, we use the blocking technique to demonstrate the effectiveness of GANs for estimating the distribution of stationary time series. Theoretically, we derive a non-asymptotic error bound for the Deep Neural Network (DNN)-based GANs estimator for the stationary distribution of the time series. Based on our theoretical analysis, we propose an algorithm for estimating the change point in time series distribution. The two main results are verified by two Monte Carlo experiments respectively, one is to estimate the joint stationary distribution of $5$-tuple samples of a 20 dimensional AR(3) model, the other is about estimating the change point at the combination of two different stationary time series. A real world empirical application to the human activity recognition dataset highlights the potential of the proposed methods.
♻ ☆ A Scalable and Efficient Iterative Method for Copying Machine Learning Classifiers
Differential replication through copying refers to the process of replicating the decision behavior of a machine learning model using another model that possesses enhanced features and attributes. This process is relevant when external constraints limit the performance of an industrial predictive system. Under such circumstances, copying enables the retention of original prediction capabilities while adapting to new demands. Previous research has focused on the single-pass implementation for copying. This paper introduces a novel sequential approach that significantly reduces the amount of computational resources needed to train or maintain a copy, leading to reduced maintenance costs for companies using machine learning models in production. The effectiveness of the sequential approach is demonstrated through experiments with synthetic and real-world datasets, showing significant reductions in time and resources, while maintaining or improving accuracy.
♻ ☆ Graph Kernels Based on Multi-scale Graph Embeddings
Graph kernels are conventional methods for computing graph similarities. However, most of the R-convolution graph kernels face two challenges: 1) They cannot compare graphs at multiple different scales, and 2) they do not consider the distributions of substructures when computing the kernel matrix. These two challenges limit their performances. To mitigate the two challenges, we propose a novel graph kernel called the Multi-scale Path-pattern Graph kernel (MPG), at the heart of which is the multi-scale path-pattern node feature map. Each element of the path-pattern node feature map is the number of occurrences of a path-pattern around a node. A path-pattern is constructed by the concatenation of all the node labels in a path of a truncated BFS tree rooted at each node. Since the path-pattern node feature map can only compare graphs at local scales, we incorporate into it the multiple different scales of the graph structure, which are captured by the truncated BFS trees of different depth. We use the Wasserstein distance to compute the similarity between the multi-scale path-pattern node feature maps of two graphs, considering the distributions of path-patterns. We empirically validate MPG on various benchmark graph datasets and demonstrate that it achieves state-of-the-art performance.
comment: 15 pages
♻ ☆ A Comprehensive Survey of Dataset Distillation
Deep learning technology has developed unprecedentedly in the last decade and has become the primary choice in many application domains. This progress is mainly attributed to a systematic collaboration in which rapidly growing computing resources encourage advanced algorithms to deal with massive data. However, it has gradually become challenging to handle the unlimited growth of data with limited computing power. To this end, diverse approaches are proposed to improve data processing efficiency. Dataset distillation, a dataset reduction method, addresses this problem by synthesizing a small typical dataset from substantial data and has attracted much attention from the deep learning community. Existing dataset distillation methods can be taxonomized into meta-learning and data matching frameworks according to whether they explicitly mimic the performance of target data. Although dataset distillation has shown surprising performance in compressing datasets, there are still several limitations such as distilling high-resolution data. This paper provides a holistic understanding of dataset distillation from multiple aspects, including distillation frameworks and algorithms, factorized dataset distillation, performance comparison, and applications. Finally, we discuss challenges and promising directions to further promote future studies on dataset distillation.
comment: 27 pages, 8 figures
♻ ☆ Koopman Operator Learning: Sharp Spectral Rates and Spurious Eigenvalues
Non-linear dynamical systems can be handily described by the associated Koopman operator, whose action evolves every observable of the system forward in time. Learning the Koopman operator from data is enabled by a number of algorithms. In this work we present nonasymptotic learning bounds for the Koopman eigenvalues and eigenfunctions estimated by two popular algorithms: Extended Dynamic Mode Decomposition (EDMD) and Reduced Rank Regression (RRR). We focus on time-reversal-invariant Markov chains, implying that the Koopman operator is self-adjoint. This includes important examples of stochastic dynamical systems, notably Langevin dynamics. Our spectral learning bounds are driven by the simultaneous control of the operator norm risk of the estimators and a metric distortion associated to the corresponding eigenfunctions. Our analysis indicates that both algorithms have similar variance, but EDMD suffers from a larger bias which might be detrimental to its learning rate. We further argue that a large metric distortion may lead to spurious eigenvalues, a phenomenon which has been empirically observed, and note that metric distortion can be estimated from data. Numerical experiments complement the theoretical findings.
comment: 8 pages, 3 figures, 6 appendices
♻ ☆ SeeGera: Self-supervised Semi-implicit Graph Variational Auto-encoders with Masking
Generative graph self-supervised learning (SSL) aims to learn node representations by reconstructing the input graph data. However, most existing methods focus on unsupervised learning tasks only and very few work has shown its superiority over the state-of-the-art graph contrastive learning (GCL) models, especially on the classification task. While a very recent model has been proposed to bridge the gap, its performance on unsupervised learning tasks is still unknown. In this paper, to comprehensively enhance the performance of generative graph SSL against other GCL models on both unsupervised and supervised learning tasks, we propose the SeeGera model, which is based on the family of self-supervised variational graph auto-encoder (VGAE). Specifically, SeeGera adopts the semi-implicit variational inference framework, a hierarchical variational framework, and mainly focuses on feature reconstruction and structure/feature masking. On the one hand, SeeGera co-embeds both nodes and features in the encoder and reconstructs both links and features in the decoder. Since feature embeddings contain rich semantic information on features, they can be combined with node embeddings to provide fine-grained knowledge for feature reconstruction. On the other hand, SeeGera adds an additional layer for structure/feature masking to the hierarchical variational framework, which boosts the model generalizability. We conduct extensive experiments comparing SeeGera with 9 other state-of-the-art competitors. Our results show that SeeGera can compare favorably against other state-of-the-art GCL methods in a variety of unsupervised and supervised learning tasks.
comment: Accepted by WebConf 2023
♻ ☆ Towards Out-of-Distribution Adversarial Robustness
Adversarial robustness continues to be a major challenge for deep learning. A core issue is that robustness to one type of attack often fails to transfer to other attacks. While prior work establishes a theoretical trade-off in robustness against different $L_p$ norms, we show that there is potential for improvement against many commonly used attacks by adopting a domain generalisation approach. Concretely, we treat each type of attack as a domain, and apply the Risk Extrapolation method (REx), which promotes similar levels of robustness against all training attacks. Compared to existing methods, we obtain similar or superior worst-case adversarial robustness on attacks seen during training. Moreover, we achieve superior performance on families or tunings of attacks only encountered at test time. On ensembles of attacks, our approach improves the accuracy from 3.4% the best existing baseline to 25.9% on MNIST, and from 16.9% to 23.5% on CIFAR10.
♻ ☆ Less is More: Understanding Word-level Textual Adversarial Attack via n-gram Frequency Descend
Word-level textual adversarial attacks have achieved striking performance in fooling natural language processing models. However, the fundamental questions of why these attacks are effective, and the intrinsic properties of the adversarial examples (AEs), are still not well understood. This work attempts to interpret textual attacks through the lens of $n$-gram frequency. Specifically, it is revealed that existing word-level attacks exhibit a strong tendency toward generation of examples with $n$-gram frequency descend ($n$-FD). Intuitively, this finding suggests a natural way to improve model robustness by training the model on the $n$-FD examples. To verify this idea, we devise a model-agnostic and gradient-free AE generation approach that relies solely on the $n$-gram frequency information, and further integrate it into the recently proposed convex hull framework for adversarial training. Surprisingly, the resultant method performs quite similarly to the original gradient-based method in terms of model robustness. These findings provide a human-understandable perspective for interpreting word-level textual adversarial attacks, and a new direction to improve model robustness.
comment: 8 pages, 4 figures. In progress
♻ ☆ Analyzing Tree Architectures in Ensembles via Neural Tangent Kernel ICLR 2023
A soft tree is an actively studied variant of a decision tree that updates splitting rules using the gradient method. Although soft trees can take various architectures, their impact is not theoretically well known. In this paper, we formulate and analyze the Neural Tangent Kernel (NTK) induced by soft tree ensembles for arbitrary tree architectures. This kernel leads to the remarkable finding that only the number of leaves at each depth is relevant for the tree architecture in ensemble learning with an infinite number of trees. In other words, if the number of leaves at each depth is fixed, the training behavior in function space and the generalization performance are exactly the same across different tree architectures, even if they are not isomorphic. We also show that the NTK of asymmetric trees like decision lists does not degenerate when they get infinitely deep. This is in contrast to the perfect binary trees, whose NTK is known to degenerate and leads to worse generalization performance for deeper trees.
comment: Accepted to ICLR 2023. arXiv admin note: text overlap with arXiv:2109.04983
♻ ☆ A Survey of Supernet Optimization and its Applications: Spatial and Temporal Optimization for Neural Architecture Search
This survey focuses on categorizing and evaluating the methods of supernet optimization in the field of Neural Architecture Search (NAS). Supernet optimization involves training a single, over-parameterized network that encompasses the search space of all possible network architectures. The survey analyses supernet optimization methods based on their approaches to spatial and temporal optimization. Spatial optimization relates to optimizing the architecture and parameters of the supernet and its subnets, while temporal optimization deals with improving the efficiency of selecting architectures from the supernet. The benefits, limitations, and potential applications of these methods in various tasks and settings, including transferability, domain generalization, and Transformer models, are also discussed.
♻ ☆ Learning Continuous Rotation Canonicalization with Radial Beam Sampling
Nearly all state of the art vision models are sensitive to image rotations. Existing methods often compensate for missing inductive biases by using augmented training data to learn pseudo-invariances. Alongside the resource demanding data inflation process, predictions often poorly generalize. The inductive biases inherent to convolutional neural networks allow for translation equivariance through kernels acting parallely to the horizontal and vertical axes of the pixel grid. This inductive bias, however, does not allow for rotation equivariance. We propose a radial beam sampling strategy along with radial kernels operating on these beams to inherently incorporate center-rotation covariance. Together with an angle distance loss, we present a radial beam-based image canonicalization model, short BIC. Our model allows for maximal continuous angle regression and canonicalizes arbitrary center-rotated input images. As a pre-processing model, this enables rotation-invariant vision pipelines with model-agnostic rotation-sensitive downstream predictions. We show that our end-to-end trained angle regressor is able to predict continuous rotation angles on several vision datasets, i.e. FashionMNIST, CIFAR10, COIL100, and LFW.
♻ ☆ Ten Lessons We Have Learned in the New "Sparseland": A Short Handbook for Sparse Neural Network Researchers
This article does not propose any novel algorithm or new hardware for sparsity. Instead, it aims to serve the "common good" for the increasingly prosperous Sparse Neural Network (SNN) research community. We attempt to summarize some most common confusions in SNNs, that one may come across in various scenarios such as paper review/rebuttal and talks - many drawn from the authors' own bittersweet experiences! We feel that doing so is meaningful and timely, since the focus of SNN research is notably shifting from traditional pruning to more diverse and profound forms of sparsity before, during, and after training. The intricate relationships between their scopes, assumptions, and approaches lead to misunderstandings, for non-experts or even experts in SNNs. In response, we summarize ten Q\&As of SNNs from many key aspects, including dense vs. sparse, unstructured sparse vs. structured sparse, pruning vs. sparse training, dense-to-sparse training vs. sparse-to-sparse training, static sparsity vs. dynamic sparsity, before-training/during-training vs. post-training sparsity, and many more. We strive to provide proper and generically applicable answers to clarify those confusions to the best extent possible. We hope our summary provides useful general knowledge for people who want to enter and engage with this exciting community; and also provides some "mind of ease" convenience for SNN researchers to explain their work in the right contexts. At the very least (and perhaps as this article's most insignificant target functionality), if you are writing/planning to write a paper or rebuttal in the field of SNNs, we hope some of our answers could help you!
♻ ☆ Projective Ranking-based GNN Evasion Attacks
Graph neural networks (GNNs) offer promising learning methods for graph-related tasks. However, GNNs are at risk of adversarial attacks. Two primary limitations of the current evasion attack methods are highlighted: (1) The current GradArgmax ignores the "long-term" benefit of the perturbation. It is faced with zero-gradient and invalid benefit estimates in certain situations. (2) In the reinforcement learning-based attack methods, the learned attack strategies might not be transferable when the attack budget changes. To this end, we first formulate the perturbation space and propose an evaluation framework and the projective ranking method. We aim to learn a powerful attack strategy then adapt it as little as possible to generate adversarial samples under dynamic budget settings. In our method, based on mutual information, we rank and assess the attack benefits of each perturbation for an effective attack strategy. By projecting the strategy, our method dramatically minimizes the cost of learning a new attack strategy when the attack budget changes. In the comparative assessment with GradArgmax and RL-S2V, the results show our method owns high attack performance and effective transferability. The visualization of our method also reveals various attack patterns in the generation of adversarial samples.
comment: Accepted by IEEE Transactions on Knowledge and Data Engineering
♻ ☆ DeepSpeed Data Efficiency: Improving Deep Learning Model Quality and Training Efficiency via Efficient Data Sampling and Routing
Recent advances on deep learning models come at the price of formidable training cost. The increasing model size is one of the root causes, but another less-emphasized fact is that data scale is actually increasing at a similar speed as model scale, and the training cost is proportional to both of them. Compared to the rapidly evolving model architecture, how to efficiently use the training data (especially for the expensive foundation model pretraining) is both less explored and difficult to realize due to the lack of a convenient framework that focus on data efficiency capabilities. To this end, we present DeepSpeed Data Efficiency, a framework that makes better use of data, increases training efficiency, and improves model quality. Specifically, we propose and combine two novel data efficiency techniques: efficient data sampling via a general curriculum learning library, and efficient data routing via a novel random layerwise token dropping technique. DeepSpeed Data Efficiency also takes extensibility, flexibility and composability into consideration, so that users can easily utilize the framework to compose multiple techniques and apply customized strategies. By applying our solution to GPT-3 1.3B and BERT-large language model pretraining, we can achieve similar model quality with up to 2x less data and 2x less time, or achieve better model quality under similar amount of data and time.
comment: Equal contribution by the first 3 authors. Code has been released as a part of https://github.com/microsoft/DeepSpeed. Part of this paper is from our previous arxiv report (arXiv:2211.11586)
♻ ☆ Rethinking Semi-Supervised Medical Image Segmentation: A Variance-Reduction Perspective
For medical image segmentation, contrastive learning is the dominant practice to improve the quality of visual representations by contrasting semantically similar and dissimilar pairs of samples. This is enabled by the observation that without accessing ground truth label, negative examples with truly dissimilar anatomical features, if sampled, can significantly improve the performance. In reality, however, these samples may come from similar anatomical features and the models may struggle to distinguish the minority tail-class samples, making the tail classes more prone to misclassification, both of which typically lead to model collapse. In this paper, we propose ARCO, a semi-supervised contrastive learning (CL) framework with stratified group sampling theory in medical image segmentation. In particular, we first propose building ARCO through the concept of variance-reduced estimation, and show that certain variance-reduction techniques are particularly beneficial in medical image segmentation tasks with extremely limited labels. Furthermore, we theoretically prove these sampling techniques are universal in variance reduction. Finally, we experimentally validate our approaches on three benchmark datasets with different label settings, and our methods consistently outperform state-of-the-art semi- and fully-supervised methods. Additionally, we augment the CL frameworks with these sampling techniques and demonstrate significant gains over previous methods. We believe our work is an important step towards semi-supervised medical image segmentation by quantifying the limitation of current self-supervision objectives for accomplishing medical image analysis tasks.
♻ ☆ A Cloud-Based Energy Management Strategy for Hybrid Electric City Bus Considering Real-Time Passenger Load Prediction
Electric city bus gains popularity in recent years for its low greenhouse gas emission, low noise level, etc. Different from a passenger car, the weight of a city bus varies significantly with different amounts of onboard passengers. After analyzing the importance of battery aging and passenger load effects on an optimal energy management strategy, this study introduces the passenger load prediction into the hybrid-electric city buses energy management problem, which is not well studied in the existing literature. The average model, Decision Tree, Gradient Boost Decision Tree, and Neural Networks models are compared in the passenger load prediction. The Gradient Boost Decision Tree model is selected due to its best accuracy and high stability. Given the predicted passenger load, a dynamic programming algorithm determines the optimal power demand for supercapacitor and battery by optimizing the battery aging and energy usage leveraging cloud techniques. Then, rule extraction is conducted on dynamic programming results, and the rule is real-time loaded to the vehicle onboard controller to handle prediction errors and uncertainties. The proposed cloud-based Dynamic Programming and rule extraction framework with the passenger load prediction show 4% and 11% lower bus operating costs in off-peak and peak hours, respectively. The operating cost by the proposed framework is less than 1% of the dynamic programming with the true passenger load information.
♻ ☆ A Data Driven Method for Multi-step Prediction of Ship Roll Motion in High Sea States
Ship roll motion in high sea states has large amplitudes and nonlinear dynamics, and its prediction is significant for operability, safety, and survivability. This paper presents a novel data-driven methodology to provide a multi-step prediction of ship roll motions in high sea states. A hybrid neural network is proposed that combines long short-term memory (LSTM) and convolutional neural network (CNN) in parallel. The motivation is to extract the nonlinear dynamic characteristics and the hydrodynamic memory information through the advantage of CNN and LSTM, respectively. For the feature selection, the time histories of motion states and wave heights are selected to involve sufficient information. Taken a scaled KCS as the study object, the ship motions in sea state 7 irregular long-crested waves are simulated and used for the validation. The results show that at least one period of roll motion can be accurately predicted. Compared with the single LSTM and CNN methods, the proposed method has better performance in predicting the amplitude of roll angles. Besides, the comparison results also demonstrate that selecting motion states and wave heights as feature space improves the prediction accuracy, verifying the effectiveness of the proposed method.
♻ ☆ Neural-network solutions to stochastic reaction networks
The stochastic reaction network in which chemical species evolve through a set of reactions is widely used to model stochastic processes in physics, chemistry and biology. To characterize the evolving joint probability distribution in the state space of species counts requires solving a system of ordinary differential equations, the chemical master equation, where the size of the counting state space increases exponentially with the type of species, making it challenging to investigate the stochastic reaction network. Here, we propose a machine-learning approach using the variational autoregressive network to solve the chemical master equation. Training the autoregressive network employs the policy gradient algorithm in the reinforcement learning framework, which does not require any data simulated in prior by another method. Different from simulating single trajectories, the approach tracks the time evolution of the joint probability distribution, and supports direct sampling of configurations and computing their normalized joint probabilities. We apply the approach to representative examples in physics and biology, and demonstrate that it accurately generates the probability distribution over time. The variational autoregressive network exhibits a plasticity in representing the multimodal distribution, cooperates with the conservation law, enables time-dependent reaction rates, and is efficient for high-dimensional reaction networks with allowing a flexible upper count limit. The results suggest a general approach to investigate stochastic reaction networks based on modern machine learning.
♻ ☆ Provable Reset-free Reinforcement Learning by No-Regret Reduction AAAI 2023
Real-world reinforcement learning (RL) is often severely limited since typical RL algorithms heavily rely on the reset mechanism to sample proper initial states. In practice, the reset mechanism is expensive to implement due to the need for human intervention or heavily engineered environments. To make learning more practical, we propose a generic no-regret reduction to systematically design reset-free RL algorithms. Our reduction turns reset-free RL into a two-player game. We show that achieving sublinear regret in this two-player game would imply learning a policy that has both sublinear performance regret and sublinear total number of resets in the original RL problem. This means that the agent eventually learns to perform optimally and avoid resets. By this reduction, we design an instantiation for linear Markov decision processes, which is the first provably correct reset-free RL algorithm to our knowledge.
comment: Full version of the paper accepted to AAAI 2023 RL4PROD Workshop. Full version of the paper also submitted to a conference and under review to be published
♻ ☆ CHiLS: Zero-Shot Image Classification with Hierarchical Label Sets
Open vocabulary models (e.g. CLIP) have shown strong performance on zero-shot classification through their ability generate embeddings for each class based on their (natural language) names. Prior work has focused on improving the accuracy of these models through prompt engineering or by incorporating a small amount of labeled downstream data (via finetuning). However, there has been little focus on improving the richness of the class names themselves, which can pose issues when class labels are coarsely-defined and uninformative. We propose Classification with Hierarchical Label Sets (or CHiLS), an alternative strategy for zero-shot classification specifically designed for datasets with implicit semantic hierarchies. CHiLS proceeds in three steps: (i) for each class, produce a set of subclasses, using either existing label hierarchies or by querying GPT-3; (ii) perform the standard zero-shot CLIP procedure as though these subclasses were the labels of interest; (iii) map the predicted subclass back to its parent to produce the final prediction. Across numerous datasets with underlying hierarchical structure, CHiLS leads to improved accuracy in situations both with and without ground-truth hierarchical information. CHiLS is simple to implement within existing CLIP pipelines and requires no additional training cost. Code is available at: https://github.com/acmi-lab/CHILS.
♻ ☆ A One-Size-Fits-All Solution to Conservative Bandit Problems
In this paper, we study a family of conservative bandit problems (CBPs) with sample-path reward constraints, i.e., the learner's reward performance must be at least as well as a given baseline at any time. We propose a One-Size-Fits-All solution to CBPs and present its applications to three encompassed problems, i.e. conservative multi-armed bandits (CMAB), conservative linear bandits (CLB) and conservative contextual combinatorial bandits (CCCB). Different from previous works which consider high probability constraints on the expected reward, we focus on a sample-path constraint on the actually received reward, and achieve better theoretical guarantees ($T$-independent additive regrets instead of $T$-dependent) and empirical performance. Furthermore, we extend the results and consider a novel conservative mean-variance bandit problem (MV-CBP), which measures the learning performance with both the expected reward and variability. For this extended problem, we provide a novel algorithm with $O(1/T)$ normalized additive regrets ($T$-independent in the cumulative form) and validate this result through empirical evaluation.
♻ ☆ GEDI: GEnerative and DIscriminative Training for Self-Supervised Learning
Self-supervised learning is a popular and powerful method for utilizing large amounts of unlabeled data, for which a wide variety of training objectives have been proposed in the literature. In this study, we perform a Bayesian analysis of state-of-the-art self-supervised learning objectives and propose a unified formulation based on likelihood learning. Our analysis suggests a simple method for integrating self-supervised learning with generative models, allowing for the joint training of these two seemingly distinct approaches. We refer to this combined framework as GEDI, which stands for GEnerative and DIscriminative training. Additionally, we demonstrate an instantiation of the GEDI framework by integrating an energy-based model with a cluster-based self-supervised learning model. Through experiments on synthetic and real-world data, including SVHN, CIFAR10, and CIFAR100, we show that GEDI outperforms existing self-supervised learning strategies in terms of clustering performance by a wide margin. We also demonstrate that GEDI can be integrated into a neural-symbolic framework to address tasks in the small data regime, where it can use logical constraints to further improve clustering and classification performance.
comment: Fixed typos/cleaned the experimental section
♻ ☆ Improving Transfer Learning with a Dual Image and Video Transformer for Multi-label Movie Trailer Genre Classification
In this paper, we study the transferability of ImageNet spatial and Kinetics spatio-temporal representations to multi-label Movie Trailer Genre Classification (MTGC). In particular, we present an extensive evaluation of the transferability of ConvNet and Transformer models pretrained on ImageNet and Kinetics to Trailers12k, a new manually-curated movie trailer dataset composed of 12,000 videos labeled with 10 different genres and associated metadata. We analyze different aspects that can influence transferability, such as frame rate, input video extension, and spatio-temporal modeling. In order to reduce the spatio-temporal structure gap between ImageNet/Kinetics and Trailers12k, we propose Dual Image and Video Transformer Architecture (DIViTA), which performs shot detection so as to segment the trailer into highly correlated clips, providing a more cohesive input for pretrained backbones and improving transferability (a 1.83% increase for ImageNet and 3.75% for Kinetics). Our results demonstrate that representations learned on either ImageNet or Kinetics are comparatively transferable to Trailers12k. Moreover, both datasets provide complementary information that can be combined to improve classification performance (a 2.91% gain compared to the top single pretraining). Interestingly, using lightweight ConvNets as pretrained backbones resulted in only a 3.46% drop in classification performance compared with the top Transformer while requiring only 11.82% of its parameters and 0.81% of its FLOPS.
♻ ☆ Reinforcing User Retention in a Billion Scale Short Video Recommender System
Recently, short video platforms have achieved rapid user growth by recommending interesting content to users. The objective of the recommendation is to optimize user retention, thereby driving the growth of DAU (Daily Active Users). Retention is a long-term feedback after multiple interactions of users and the system, and it is hard to decompose retention reward to each item or a list of items. Thus traditional point-wise and list-wise models are not able to optimize retention. In this paper, we choose reinforcement learning methods to optimize the retention as they are designed to maximize the long-term performance. We formulate the problem as an infinite-horizon request-based Markov Decision Process, and our objective is to minimize the accumulated time interval of multiple sessions, which is equal to improving the app open frequency and user retention. However, current reinforcement learning algorithms can not be directly applied in this setting due to uncertainty, bias, and long delay time incurred by the properties of user retention. We propose a novel method, dubbed RLUR, to address the aforementioned challenges. Both offline and live experiments show that RLUR can significantly improve user retention. RLUR has been fully launched in Kuaishou app for a long time, and achieves consistent performance improvement on user retention and DAU.
♻ ☆ Don't Blame the Annotator: Bias Already Starts in the Annotation Instructions EACL 2023
In recent years, progress in NLU has been driven by benchmarks. These benchmarks are typically collected by crowdsourcing, where annotators write examples based on annotation instructions crafted by dataset creators. In this work, we hypothesize that annotators pick up on patterns in the crowdsourcing instructions, which bias them to write many similar examples that are then over-represented in the collected data. We study this form of bias, termed instruction bias, in 14 recent NLU benchmarks, showing that instruction examples often exhibit concrete patterns, which are propagated by crowdworkers to the collected data. This extends previous work (Geva et al., 2019) and raises a new concern of whether we are modeling the dataset creator's instructions, rather than the task. Through a series of experiments, we show that, indeed, instruction bias can lead to overestimation of model performance, and that models struggle to generalize beyond biases originating in the crowdsourcing instructions. We further analyze the influence of instruction bias in terms of pattern frequency and model size, and derive concrete recommendations for creating future NLU benchmarks.
comment: Accepted to EACL 2023
♻ ☆ Co-Imitation: Learning Design and Behaviour by Imitation AAAI-23
The co-adaptation of robots has been a long-standing research endeavour with the goal of adapting both body and behaviour of a system for a given task, inspired by the natural evolution of animals. Co-adaptation has the potential to eliminate costly manual hardware engineering as well as improve the performance of systems. The standard approach to co-adaptation is to use a reward function for optimizing behaviour and morphology. However, defining and constructing such reward functions is notoriously difficult and often a significant engineering effort. This paper introduces a new viewpoint on the co-adaptation problem, which we call co-imitation: finding a morphology and a policy that allow an imitator to closely match the behaviour of a demonstrator. To this end we propose a co-imitation methodology for adapting behaviour and morphology by matching state distributions of the demonstrator. Specifically, we focus on the challenging scenario with mismatched state- and action-spaces between both agents. We find that co-imitation increases behaviour similarity across a variety of tasks and settings, and demonstrate co-imitation by transferring human walking, jogging and kicking skills onto a simulated humanoid.
comment: 14 pages, 11 figures, accepted for AAAI-23
♻ ☆ This Intestine Does Not Exist: Multiscale Residual Variational Autoencoder for Realistic Wireless Capsule Endoscopy Image Generation
Medical image synthesis has emerged as a promising solution to address the limited availability of annotated medical data needed for training machine learning algorithms in the context of image-based Clinical Decision Support (CDS) systems. To this end, Generative Adversarial Networks (GANs) have been mainly applied to support the algorithm training process by generating synthetic images for data augmentation. However, in the field of Wireless Capsule Endoscopy (WCE), the limited content diversity and size of existing publicly available annotated datasets, adversely affect both the training stability and synthesis performance of GANs. Aiming to a viable solution for WCE image synthesis, a novel Variational Autoencoder architecture is proposed, namely "This Intestine Does not Exist" (TIDE). The proposed architecture comprises multiscale feature extraction convolutional blocks and residual connections, which enable the generation of high-quality and diverse datasets even with a limited number of training images. Contrary to the current approaches, which are oriented towards the augmentation of the available datasets, this study demonstrates that using TIDE, real WCE datasets can be fully substituted by artificially generated ones, without compromising classification performance. Furthermore, qualitative and user evaluation studies by experienced WCE specialists, validate from a medical viewpoint that both the normal and abnormal WCE images synthesized by TIDE are sufficiently realistic.
comment: 10 pages
♻ ☆ Fuzzy Expert Systems for Prediction of ICU Admission in Patients with COVID-19
The pandemic COVID-19 disease has had a dramatic impact on almost all countries around the world so that many hospitals have been overwhelmed with Covid-19 cases. As medical resources are limited, deciding on the proper allocation of these resources is a very crucial issue. Besides, uncertainty is a major factor that can affect decisions, especially in medical fields. To cope with this issue, we use fuzzy logic (FL) as one of the most suitable methods in modeling systems with high uncertainty and complexity. We intend to make use of the advantages of FL in decisions on cases that need to treat in ICU. In this study, an interval type-2 fuzzy expert system is proposed for prediction of ICU admission in COVID-19 patients. For this prediction task, we also developed an adaptive neuro-fuzzy inference system (ANFIS). Finally, the results of these fuzzy systems are compared to some well-known classification methods such as Naive Bayes (NB), Case-Based Reasoning (CBR), Decision Tree (DT), and K Nearest Neighbor (KNN). The results show that the type-2 fuzzy expert system and ANFIS models perform competitively in terms of accuracy and F-measure compared to the other system modeling techniques.
♻ ☆ MetaQA: Combining Expert Agents for Multi-Skill Question Answering EACL 2023
The recent explosion of question answering (QA) datasets and models has increased the interest in the generalization of models across multiple domains and formats by either training on multiple datasets or by combining multiple models. Despite the promising results of multi-dataset models, some domains or QA formats may require specific architectures, and thus the adaptability of these models might be limited. In addition, current approaches for combining models disregard cues such as question-answer compatibility. In this work, we propose to combine expert agents with a novel, flexible, and training-efficient architecture that considers questions, answer predictions, and answer-prediction confidence scores to select the best answer among a list of answer candidates. Through quantitative and qualitative experiments we show that our model i) creates a collaboration between agents that outperforms previous multi-agent and multi-dataset approaches in both in-domain and out-of-domain scenarios, ii) is highly data-efficient to train, and iii) can be adapted to any QA format. We release our code and a dataset of answer predictions from expert agents for 16 QA datasets to foster future developments of multi-agent systems on https://github.com/UKPLab/MetaQA.
comment: Accepted at EACL 2023
♻ ☆ Backdoor Attacks in Peer-to-Peer Federated Learning
We study backdoor attacks in peer-to-peer federated learning systems on different graph topologies and datasets. We show that only 5% attacker nodes are sufficient to perform a backdoor attack with 42% attack success without decreasing the accuracy on clean data by more than 2%. We also demonstrate that the attack can be amplified by the attacker crashing a small number of nodes. We evaluate defenses proposed in the context of centralized federated learning and show they are ineffective in peer-to-peer settings. Finally, we propose a defense that mitigates the attacks by applying different clipping norms to the model updates received from peers and local model trained by a node.
♻ ☆ Interpolation for Robust Learning: Data Augmentation on Geodesics
We propose to study and promote the robustness of a model as per its performance through the interpolation of training data distributions. Specifically, (1) we augment the data by finding the worst-case Wasserstein barycenter on the geodesic connecting subpopulation distributions of different categories. (2) We regularize the model for smoother performance on the continuous geodesic path connecting subpopulation distributions. (3) Additionally, we provide a theoretical guarantee of robustness improvement and investigate how the geodesic location and the sample size contribute, respectively. Experimental validations of the proposed strategy on four datasets, including CIFAR-100 and ImageNet, establish the efficacy of our method, e.g., our method improves the baselines' certifiable robustness on CIFAR10 up to $7.7\%$, with $16.8\%$ on empirical robustness on CIFAR-100. Our work provides a new perspective of model robustness through the lens of Wasserstein geodesic-based interpolation with a practical off-the-shelf strategy that can be combined with existing robust training methods.
comment: 33 pages, 3 figures, 18 tables
♻ ☆ (Private) Kernelized Bandits with Distributed Biased Feedback
In this paper, we study kernelized bandits with distributed biased feedback. This problem is motivated by several real-world applications (such as dynamic pricing, cellular network configuration, and policy making), where users from a large population contribute to the reward of the action chosen by a central entity, but it is difficult to collect feedback from all users. Instead, only biased feedback (due to user heterogeneity) from a subset of users may be available. In addition to such partial biased feedback, we are also faced with two practical challenges due to communication cost and computation complexity. To tackle these challenges, we carefully design a new \emph{distributed phase-then-batch-based elimination (\texttt{DPBE})} algorithm, which samples users in phases for collecting feedback to reduce the bias and employs \emph{maximum variance reduction} to select actions in batches within each phase. By properly choosing the phase length, the batch size, and the confidence width used for eliminating suboptimal actions, we show that \texttt{DPBE} achieves a sublinear regret of $\tilde{O}(T^{1-\alpha/2}+\sqrt{\gamma_T T})$, where $\alpha\in (0,1)$ is the user-sampling parameter one can tune. Moreover, \texttt{DPBE} can significantly reduce both communication cost and computation complexity in distributed kernelized bandits, compared to some variants of the state-of-the-art algorithms (originally developed for standard kernelized bandits). Furthermore, by incorporating various \emph{differential privacy} models (including the central, local, and shuffle models), we generalize \texttt{DPBE} to provide privacy guarantees for users participating in the distributed learning process. Finally, we conduct extensive simulations to validate our theoretical results and evaluate the empirical performance.
comment: This work has been accepted by ACM SIGMETRICS 2023
Multimedia 4
☆ Revisiting Pre-training in Audio-Visual Learning
Pre-training technique has gained tremendous success in enhancing model performance on various tasks, but found to perform worse than training from scratch in some uni-modal situations. This inspires us to think: are the pre-trained models always effective in the more complex multi-modal scenario, especially for the heterogeneous modalities such as audio and visual ones? We find that the answer is No. Specifically, we explore the effects of pre-trained models on two audio-visual learning scenarios: cross-modal initialization and multi-modal joint learning. When cross-modal initialization is applied, the phenomena of "dead channel" caused by abnormal Batchnorm parameters hinders the utilization of model capacity. Thus, we propose Adaptive Batchnorm Re-initialization (ABRi) to better exploit the capacity of pre-trained models for target tasks. In multi-modal joint learning, we find a strong pre-trained uni-modal encoder would bring negative effects on the encoder of another modality. To alleviate such problem, we introduce a two-stage Fusion Tuning strategy, taking better advantage of the pre-trained knowledge while making the uni-modal encoders cooperate with an adaptive masking method. The experiment results show that our methods could further exploit pre-trained models' potential and boost performance in audio-visual learning.
☆ Online Misinformation Video Detection: A Survey
With information consumption via online video streaming becoming increasingly popular, misinformation video poses a new threat to the health of the online information ecosystem. Though previous studies have made much progress in detecting misinformation in text and image formats, video-based misinformation brings new and unique challenges to automatic detection systems: 1) high information heterogeneity brought by various modalities, 2) blurred distinction between misleading video manipulation and ubiquitous artistic video editing, and 3) new patterns of misinformation propagation due to the dominant role of recommendation systems on online video platforms. To facilitate research on this challenging task, we conduct this survey to present advances in misinformation video detection research. We first analyze and characterize the misinformation video from three levels including signals, semantics, and intents. Based on the characterization, we systematically review existing works for detection from features of various modalities to techniques for clue integration. We also introduce existing resources including representative datasets and widely used tools. Besides summarizing existing studies, we discuss related areas and outline open issues and future directions to encourage and guide more research on misinformation video detection. Our corresponding public repository is available at https://github.com/ICTMCG/Awesome-Misinfo-Video-Detection.
comment: 10 pages, 2 figures
♻ ☆ Video compression dataset and benchmark of learning-based video-quality metrics
Video-quality measurement is a critical task in video processing. Nowadays, many implementations of new encoding standards - such as AV1, VVC, and LCEVC - use deep-learning-based decoding algorithms with perceptual metrics that serve as optimization objectives. But investigations of the performance of modern video- and image-quality metrics commonly employ videos compressed using older standards, such as AVC. In this paper, we present a new benchmark for video-quality metrics that evaluates video compression. It is based on a new dataset consisting of about 2,500 streams encoded using different standards, including AVC, HEVC, AV1, VP9, and VVC. Subjective scores were collected using crowdsourced pairwise comparisons. The list of evaluated metrics includes recent ones based on machine learning and neural networks. The results demonstrate that new no-reference metrics exhibit a high correlation with subjective quality and approach the capability of top full-reference metrics.
comment: 10 pages, 4 figures, 6 tables, 1 supplementary material
♻ ☆ Forgetting to Remember: A Scalable Incremental Learning Framework for Cross-Task Blind Image Quality Assessment
Recent years have witnessed the great success of blind image quality assessment (BIQA) in various task-specific scenarios, which present invariable distortion types and evaluation criteria. However, due to the rigid structure and learning framework, they cannot apply to the cross-task BIQA scenario, where the distortion types and evaluation criteria keep changing in practical applications. This paper proposes a scalable incremental learning framework (SILF) that could sequentially conduct BIQA across multiple evaluation tasks with limited memory capacity. More specifically, we develop a dynamic parameter isolation strategy to sequentially update the task-specific parameter subsets, which are non-overlapped with each other. Each parameter subset is temporarily settled to Remember one evaluation preference toward its corresponding task, and the previously settled parameter subsets can be adaptively reused in the following BIQA to achieve better performance based on the task relevance. To suppress the unrestrained expansion of memory capacity in sequential tasks learning, we develop a scalable memory unit by gradually and selectively pruning unimportant neurons from previously settled parameter subsets, which enable us to Forget part of previous experiences and free the limited memory capacity for adapting to the emerging new tasks. Extensive experiments on eleven IQA datasets demonstrate that our proposed method significantly outperforms the other state-of-the-art methods in cross-task BIQA. The source code of the proposed method is available at https://github.com/maruiperfect/SILF.
Computation and Language 40
☆ Erasure of Unaligned Attributes from Neural Representations
We present the Assignment-Maximization Spectral Attribute removaL (AMSAL) algorithm, which aims at removing information from neural representations when the information to be erased is implicit rather than directly being aligned to each input example. Our algorithm works by alternating between two steps. In one, it finds an assignment of the input representations to the information to be erased, and in the other, it creates projections of both the input representations and the information to be erased into a joint latent space. We test our algorithm on an extensive array of datasets, including a Twitter dataset with multiple guarded attributes, the BiasBios dataset and the BiasBench benchmark. The latter benchmark includes four datasets with various types of protected attributes. Our results demonstrate that bias can often be removed in our setup. We also discuss the limitations of our approach when there is a strong entanglement between the main task and the information to be erased.
comment: Accepted to Transactions of the Association for Computational Linguistics, 22 pages (pre-MIT Press publication version)
☆ LoFT: Enhancing Faithfulness and Diversity for Table-to-Text Generation via Logic Form Control EACL 2023
Logical Table-to-Text (LT2T) generation is tasked with generating logically faithful sentences from tables. There currently exists two challenges in the field: 1) Faithfulness: how to generate sentences that are factually correct given the table content; 2) Diversity: how to generate multiple sentences that offer different perspectives on the table. This work proposes LoFT, which utilizes logic forms as fact verifiers and content planners to control LT2T generation. Experimental results on the LogicNLG dataset demonstrate that LoFT is the first model that addresses unfaithfulness and lack of diversity issues simultaneously. Our code is publicly available at https://github.com/Yale-LILY/LoFT.
comment: Accepted at EACL 2023 as a short paper
☆ Controllable Lexical Simplification for English
Fine-tuning Transformer-based approaches have recently shown exciting results on sentence simplification task. However, so far, no research has applied similar approaches to the Lexical Simplification (LS) task. In this paper, we present ConLS, a Controllable Lexical Simplification system fine-tuned with T5 (a Transformer-based model pre-trained with a BERT-style approach and several other tasks). The evaluation results on three datasets (LexMTurk, BenchLS, and NNSeval) have shown that our model performs comparable to LSBert (the current state-of-the-art) and even outperforms it in some cases. We also conducted a detailed comparison on the effectiveness of control tokens to give a clear view of how each token contributes to the model.
☆ Findings of the TSAR-2022 Shared Task on Multilingual Lexical Simplification
We report findings of the TSAR-2022 shared task on multilingual lexical simplification, organized as part of the Workshop on Text Simplification, Accessibility, and Readability TSAR-2022 held in conjunction with EMNLP 2022. The task called the Natural Language Processing research community to contribute with methods to advance the state of the art in multilingual lexical simplification for English, Portuguese, and Spanish. A total of 14 teams submitted the results of their lexical simplification systems for the provided test data. Results of the shared task indicate new benchmarks in Lexical Simplification with English lexical simplification quantitative results noticeably higher than those obtained for Spanish and (Brazilian) Portuguese.
☆ Guide the Learner: Controlling Product of Experts Debiasing Method Based on Token Attribution Similarities EACL 2023
Several proposals have been put forward in recent years for improving out-of-distribution (OOD) performance through mitigating dataset biases. A popular workaround is to train a robust model by re-weighting training examples based on a secondary biased model. Here, the underlying assumption is that the biased model resorts to shortcut features. Hence, those training examples that are correctly predicted by the biased model are flagged as being biased and are down-weighted during the training of the main model. However, assessing the importance of an instance merely based on the predictions of the biased model may be too naive. It is possible that the prediction of the main model can be derived from another decision-making process that is distinct from the behavior of the biased model. To circumvent this, we introduce a fine-tuning strategy that incorporates the similarity between the main and biased model attribution scores in a Product of Experts (PoE) loss function to further improve OOD performance. With experiments conducted on natural language inference and fact verification benchmarks, we show that our method improves OOD results while maintaining in-distribution (ID) performance.
comment: Accepted to EACL 2023 (main conference)
☆ Migration Reframed? A multilingual analysis on the stance shift in Europe during the Ukrainian crisis
The war in Ukraine seems to have positively changed the attitude toward the critical societal topic of migration in Europe -- at least towards refugees from Ukraine. We investigate whether this impression is substantiated by how the topic is reflected in online news and social media, thus linking the representation of the issue on the Web to its perception in society. For this purpose, we combine and adapt leading-edge automatic text processing for a novel multilingual stance detection approach. Starting from 5.5M Twitter posts published by 565 European news outlets in one year, beginning September 2021, plus replies, we perform a multilingual analysis of migration-related media coverage and associated social media interaction for Europe and selected European countries. The results of our analysis show that there is actually a reframing of the discussion illustrated by the terminology change, e.g., from "migrant" to "refugee", often even accentuated with phrases such as "real refugees". However, concerning a stance shift in public perception, the picture is more diverse than expected. All analyzed cases show a noticeable temporal stance shift around the start of the war in Ukraine. Still, there are apparent national differences in the size and stability of this shift.
comment: To be published in The Web Conference 2023
☆ Coherence and Diversity through Noise: Self-Supervised Paraphrase Generation via Structure-Aware Denoising
In this paper, we propose SCANING, an unsupervised framework for paraphrasing via controlled noise injection. We focus on the novel task of paraphrasing algebraic word problems having practical applications in online pedagogy as a means to reduce plagiarism as well as ensure understanding on the part of the student instead of rote memorization. This task is more complex than paraphrasing general-domain corpora due to the difficulty in preserving critical information for solution consistency of the paraphrased word problem, managing the increased length of the text and ensuring diversity in the generated paraphrase. Existing approaches fail to demonstrate adequate performance on at least one, if not all, of these facets, necessitating the need for a more comprehensive solution. To this end, we model the noising search space as a composition of contextual and syntactic aspects and sample noising functions consisting of either one or both aspects. This allows for learning a denoising function that operates over both aspects and produces semantically equivalent and syntactically diverse outputs through grounded noise injection. The denoising function serves as a foundation for learning a paraphrasing function which operates solely in the input-paraphrase space without carrying any direct dependency on noise. We demonstrate SCANING considerably improves performance in terms of both semantic preservation and producing diverse paraphrases through extensive automated and manual evaluation across 4 datasets.
comment: 12 pages (main}; 22 pages in total
☆ Languages are Rewards: Hindsight Finetuning using Human Feedback
Learning from human preferences is important for language models to be helpful and useful for humans, and to align with human and social values. Existing works focus on supervised finetuning of pretrained models, based on curated model generations that are preferred by human labelers. Such works have achieved remarkable successes in understanding and following instructions (e.g., InstructGPT, ChatGPT, etc). However, to date, a key limitation of supervised finetuning is that it cannot learn from negative ratings; models are only trained on positive-rated data, which makes it data inefficient. Because collecting human feedback data is both time consuming and expensive, it is vital for the model to learn from all feedback, akin to the remarkable ability of humans to learn from diverse feedback. In this work, we propose a novel technique called Hindsight Finetuning for making language models learn from diverse human feedback. In fact, our idea is motivated by how humans learn from hindsight experience. We condition the model on a sequence of model generations paired with hindsight feedback, and finetune the model to predict the most preferred output. By doing so, models can learn to identify and correct negative attributes or errors. Applying the method to GPT-J, we observe that it significantly improves results on summarization and dialogue tasks using the same amount of human feedback.
☆ Evolution of grammatical forms: some quantitative approaches
Grammatical forms are said to evolve via two main mechanisms. These are, respectively, the `descent' mechanism, where current forms can be seen to have descended (albeit with occasional modifications) from their roots in ancient languages, and the `contact' mechanism, where evolution in a given language occurs via borrowing from other languages with which it is in contact. We use ideas and concepts from statistical physics to formulate a series of static and dynamical models which illustrate these issues in general terms. The static models emphasise the relative numbers of rules and exceptions, while the dynamical models focus on the emergence of exceptional forms. These unlikely survivors among various competing grammatical forms are winners against the odds. Our analysis suggests that they emerge when the influence of neighbouring languages exceeds the generic tendency towards regularisation within individual languages.
comment: 19 pages, 10 figures
☆ Less is More: Understanding Word-level Textual Adversarial Attack via n-gram Frequency Descend
Word-level textual adversarial attacks have achieved striking performance in fooling natural language processing models. However, the fundamental questions of why these attacks are effective, and the intrinsic properties of the adversarial examples (AEs), are still not well understood. This work attempts to interpret textual attacks through the lens of $n$-gram frequency. Specifically, it is revealed that existing word-level attacks exhibit a strong tendency toward generation of examples with $n$-gram frequency descend ($n$-FD). Intuitively, this finding suggests a natural way to improve model robustness by training the model on the $n$-FD examples. To verify this idea, we devise a model-agnostic and gradient-free AE generation approach that relies solely on the $n$-gram frequency information, and further integrate it into the recently proposed convex hull framework for adversarial training. Surprisingly, the resultant method performs quite similarly to the original gradient-based method in terms of model robustness. These findings provide a human-understandable perspective for interpreting word-level textual adversarial attacks, and a new direction to improve model robustness.
comment: 8 pages, 4 figures. In progress
☆ Data Selection for Language Models via Importance Resampling
Selecting a suitable training dataset is crucial for both general-domain (e.g., GPT-3) and domain-specific (e.g., Codex) language models (LMs). We formalize this data selection problem as selecting a subset of a large raw unlabeled dataset to match a desired target distribution, given some unlabeled target samples. Due to the large scale and dimensionality of the raw text data, existing methods use simple heuristics to select data that are similar to a high-quality reference corpus (e.g., Wikipedia), or leverage experts to manually curate data. Instead, we extend the classic importance resampling approach used in low-dimensions for LM data selection. Crucially, we work in a reduced feature space to make importance weight estimation tractable over the space of text. To determine an appropriate feature space, we first show that KL reduction, a data metric that measures the proximity between selected data and the target in a feature space, has high correlation with average accuracy on 8 downstream tasks (r=0.89) when computed with simple n-gram features. From this observation, we present Data Selection with Importance Resampling (DSIR), an efficient and scalable algorithm that estimates importance weights in a reduced feature space (e.g., n-gram features in our instantiation) and selects data with importance resampling according to these weights. When training general-domain models (target is Wikipedia + books), DSIR improves over random selection and heuristic filtering baselines by 2--2.5% on the GLUE benchmark. When performing continued pretraining towards a specific domain, DSIR performs comparably to expert curated data across 8 target distributions.
☆ Protecting Language Generation Models via Invisible Watermarking
Language generation models have been an increasingly powerful enabler for many applications. Many such models offer free or affordable API access, which makes them potentially vulnerable to model extraction attacks through distillation. To protect intellectual property (IP) and ensure fair use of these models, various techniques such as lexical watermarking and synonym replacement have been proposed. However, these methods can be nullified by obvious countermeasures such as "synonym randomization". To address this issue, we propose GINSEW, a novel method to protect text generation models from being stolen through distillation. The key idea of our method is to inject secret signals into the probability vector of the decoding steps for each target token. We can then detect the secret message by probing a suspect model to tell if it is distilled from the protected one. Experimental results show that GINSEW can effectively identify instances of IP infringement with minimal impact on the generation quality of protected APIs. Our method demonstrates an absolute improvement of 19 to 29 points on mean average precision (mAP) in detecting suspects compared to previous methods against watermark removal attacks.
☆ It's about Time: Rethinking Evaluation on Rumor Detection Benchmarks using Chronological Splits EACL 2023
New events emerge over time influencing the topics of rumors in social media. Current rumor detection benchmarks use random splits as training, development and test sets which typically results in topical overlaps. Consequently, models trained on random splits may not perform well on rumor classification on previously unseen topics due to the temporal concept drift. In this paper, we provide a re-evaluation of classification models on four popular rumor detection benchmarks considering chronological instead of random splits. Our experimental results show that the use of random splits can significantly overestimate predictive performance across all datasets and models. Therefore, we suggest that rumor detection models should always be evaluated using chronological splits for minimizing topical overlaps.
comment: Accepted at EACL 2023 Findings
☆ Techniques to Improve Neural Math Word Problem Solvers
Developing automatic Math Word Problem (MWP) solvers is a challenging task that demands the ability of understanding and mathematical reasoning over the natural language. Recent neural-based approaches mainly encode the problem text using a language model and decode a mathematical expression over quantities and operators iteratively. Note the problem text of a MWP consists of a context part and a question part, a recent work finds these neural solvers may only perform shallow pattern matching between the context text and the golden expression, where question text is not well used. Meanwhile, existing decoding processes fail to enforce the mathematical laws into the design, where the representations for mathematical equivalent expressions are different. To address these two issues, we propose a new encoder-decoder architecture that fully leverages the question text and preserves step-wise commutative law. Besides generating quantity embeddings, our encoder further encodes the question text and uses it to guide the decoding process. At each step, our decoder uses Deep Sets to compute expression representations so that these embeddings are invariant under any permutation of quantities. Experiments on four established benchmarks demonstrate that our framework outperforms state-of-the-art neural MWP solvers, showing the effectiveness of our techniques. We also conduct a detailed analysis of the results to show the limitations of our approach and further discuss the potential future work. Code is available at https://github.com/sophistz/Question-Aware-Deductive-MWP.
☆ Context-Gloss Augmentation for Improving Arabic Target Sense Verification
Arabic language lacks semantic datasets and sense inventories. The most common semantically-labeled dataset for Arabic is the ArabGlossBERT, a relatively small dataset that consists of 167K context-gloss pairs (about 60K positive and 107K negative pairs), collected from Arabic dictionaries. This paper presents an enrichment to the ArabGlossBERT dataset, by augmenting it using (Arabic-English-Arabic) machine back-translation. Augmentation increased the dataset size to 352K pairs (149K positive and 203K negative pairs). We measure the impact of augmentation using different data configurations to fine-tune BERT on target sense verification (TSV) task. Overall, the accuracy ranges between 78% to 84% for different data configurations. Although our approach performed at par with the baseline, we did observe some improvements for some POS tags in some experiments. Furthermore, our fine-tuned models are trained on a larger dataset covering larger vocabulary and contexts. We provide an in-depth analysis of the accuracy for each part-of-speech (POS).
☆ Efficient and Flexible Topic Modeling using Pretrained Embeddings and Bag of Sentences
Pre-trained language models have led to a new state-of-the-art in many NLP tasks. However, for topic modeling, statistical generative models such as LDA are still prevalent, which do not easily allow incorporating contextual word vectors. They might yield topics that do not align very well with human judgment. In this work, we propose a novel topic modeling and inference algorithm. We suggest a bag of sentences (BoS) approach using sentences as the unit of analysis. We leverage pre-trained sentence embeddings by combining generative process models with clustering. We derive a fast inference algorithm based on expectation maximization, hard assignments, and an annealing process. Our evaluation shows that our method yields state-of-the art results with relatively little computational demands. Our methods is more flexible compared to prior works leveraging word embeddings, since it provides the possibility to customize topic-document distributions using priors. Code is at \url{https://github.com/JohnTailor/BertSenClu}.
☆ APAM: Adaptive Pre-training and Adaptive Meta Learning in Language Model for Noisy Labels and Long-tailed Learning
Practical natural language processing (NLP) tasks are commonly long-tailed with noisy labels. Those problems challenge the generalization and robustness of complex models such as Deep Neural Networks (DNNs). Some commonly used resampling techniques, such as oversampling or undersampling, could easily lead to overfitting. It is growing popular to learn the data weights leveraging a small amount of metadata. Besides, recent studies have shown the advantages of self-supervised pre-training, particularly to the under-represented data. In this work, we propose a general framework to handle the problem of both long-tail and noisy labels. The model is adapted to the domain of problems in a contrastive learning manner. The re-weighting module is a feed-forward network that learns explicit weighting functions and adapts weights according to metadata. The framework further adapts weights of terms in the loss function through a combination of the polynomial expansion of cross-entropy loss and focal loss. Our extensive experiments show that the proposed framework consistently outperforms baseline methods. Lastly, our sensitive analysis emphasizes the capability of the proposed framework to handle the long-tailed problem and mitigate the negative impact of noisy labels.
☆ A Categorical Archive of ChatGPT Failures
Large language models have been demonstrated to be valuable in different fields. ChatGPT, developed by OpenAI, has been trained using massive amounts of data and simulates human conversation by comprehending context and generating appropriate responses. It has garnered significant attention due to its ability to effectively answer a broad range of human inquiries, with fluent and comprehensive answers surpassing prior public chatbots in both security and usefulness. However, a comprehensive analysis of ChatGPT's failures is lacking, which is the focus of this study. Ten categories of failures, including reasoning, factual errors, math, coding, and bias, are presented and discussed. The risks, limitations, and societal implications of ChatGPT are also highlighted. The goal of this study is to assist researchers and developers in enhancing future language models and chatbots.
♻ ☆ Improving Few-Shot Generalization by Exploring and Exploiting Auxiliary Data
Few-shot learning involves learning an effective model from only a few labeled datapoints. The use of a small training set makes it difficult to avoid overfitting but also makes few-shot learning applicable to many important real-world settings. In this work, we focus on Few-shot Learning with Auxiliary Data (FLAD), a training paradigm that assumes access to auxiliary data during few-shot learning in hopes of improving generalization. Introducing auxiliary data during few-shot learning leads to essential design choices where hand-designed heuristics can lead to sub-optimal performance. In this work, we focus on automated sampling strategies for FLAD and relate them to the explore-exploit dilemma that is central in multi-armed bandit settings. Based on this connection we propose two algorithms -- EXP3-FLAD and UCB1-FLAD -- and compare them with methods that either explore or exploit, finding that the combination of exploration and exploitation is crucial. Using our proposed algorithms to train T5 yields a 9% absolute improvement over the explicitly multi-task pre-trained T0 model across 11 datasets.
comment: 19 pages, 7 figures, code available at https://github.com/alon-albalak/FLAD
♻ ☆ Wolfies at SemEval-2022 Task 8: Feature extraction pipeline with transformers for Multi-lingual news article similarity
This work is about finding the similarity between a pair of news articles. There are seven different objective similarity metrics provided in the dataset for each pair and the news articles are in multiple different languages. On top of the pre-trained embedding model, we calculated cosine similarity for baseline results and feed-forward neural network was then trained on top of it to improve the results. We also built separate pipelines for each similarity metric for feature extraction. We could see significant improvement from baseline results using feature extraction and feed-forward neural network.
♻ ☆ DiscoScore: Evaluating Text Generation with BERT and Discourse Coherence EACL2023
Recently, there has been a growing interest in designing text generation systems from a discourse coherence perspective, e.g., modeling the interdependence between sentences. Still, recent BERT-based evaluation metrics are weak in recognizing coherence, and thus are not reliable in a way to spot the discourse-level improvements of those text generation systems. In this work, we introduce DiscoScore, a parametrized discourse metric, which uses BERT to model discourse coherence from different perspectives, driven by Centering theory. Our experiments encompass 16 non-discourse and discourse metrics, including DiscoScore and popular coherence models, evaluated on summarization and document-level machine translation (MT). We find that (i) the majority of BERT-based metrics correlate much worse with human rated coherence than early discourse metrics, invented a decade ago; (ii) the recent state-of-the-art BARTScore is weak when operated at system level -- which is particularly problematic as systems are typically compared in this manner. DiscoScore, in contrast, achieves strong system-level correlation with human ratings, not only in coherence but also in factual consistency and other aspects, and surpasses BARTScore by over 10 correlation points on average. Further, aiming to understand DiscoScore, we provide justifications to the importance of discourse coherence for evaluation metrics, and explain the superiority of one variant over another. Our code is available at \url{https://github.com/AIPHES/DiscoScore}.
comment: EACL2023 Camera Ready
♻ ☆ Locally Typical Sampling ACL 2022
Today's probabilistic language generators fall short when it comes to producing coherent and fluent text despite the fact that the underlying models perform well under standard metrics, e.g., perplexity. This discrepancy has puzzled the language generation community for the last few years. In this work, we posit that the abstraction of natural language generation as a discrete stochastic process--which allows for an information-theoretic analysis--can provide new insights into the behavior of probabilistic language generators, e.g., why high-probability texts can be dull or repetitive. Humans use language as a means of communicating information, aiming to do so in a simultaneously efficient and error-minimizing manner; in fact, psycholinguistics research suggests humans choose each word in a string with this subconscious goal in mind. We formally define the set of strings that meet this criterion: those for which each word has an information content close to the expected information content, i.e., the conditional entropy of our model. We then propose a simple and efficient procedure for enforcing this criterion when generating from probabilistic models, which we call locally typical sampling. Automatic and human evaluations show that, in comparison to nucleus and top-k sampling, locally typical sampling offers competitive performance (in both abstractive summarization and story generation) in terms of quality while consistently reducing degenerate repetitions.
comment: TACL 2022
♻ ☆ Reasoning Through Memorization: Nearest Neighbor Knowledge Graph Embeddings
Previous knowledge graph embedding approaches usually map entities to representations and utilize score functions to predict the target entities, yet they typically struggle to reason rare or emerging unseen entities. In this paper, we propose kNN-KGE, a new knowledge graph embedding approach with pre-trained language models, by linearly interpolating its entity distribution with k-nearest neighbors. We compute the nearest neighbors based on the distance in the entity embedding space from the knowledge store. Our approach can allow rare or emerging entities to be memorized explicitly rather than implicitly in model parameters. Experimental results demonstrate that our approach can improve inductive and transductive link prediction results and yield better performance for low-resource settings with only a few triples, which might be easier to reason via explicit memory. Code is available at https://github.com/zjunlp/KNN-KG.
comment: Work in progress
♻ ☆ FewShotTextGCN: K-hop neighborhood regularization for few-shot learning on graphs EACL 2023
We present FewShotTextGCN, a novel method designed to effectively utilize the properties of word-document graphs for improved learning in low-resource settings. We introduce K-hop Neighbourhood Regularization, a regularizer for heterogeneous graphs, and show that it stabilizes and improves learning when only a few training samples are available. We furthermore propose a simplification in the graph-construction method, which results in a graph that is $\sim$7 times less dense and yields better performance in little-resource settings while performing on par with the state of the art in high-resource settings. Finally, we introduce a new variant of Adaptive Pseudo-Labeling tailored for word-document graphs. When using as little as 20 samples for training, we outperform a strong TextGCN baseline with 17% in absolute accuracy on average over eight languages. We demonstrate that our method can be applied to document classification without any language model pretraining on a wide range of typologically diverse languages while performing on par with large pretrained language models.
comment: 8 pages, 4 figures, EACL 2023
♻ ☆ Data Augmentation for Automated Essay Scoring using Transformer Models
Automated essay scoring is one of the most important problem in Natural Language Processing. It has been explored for a number of years, and it remains partially solved. In addition to its economic and educational usefulness, it presents research problems. Transfer learning has proved to be beneficial in NLP. Data augmentation techniques have also helped build state-of-the-art models for automated essay scoring. Many works in the past have attempted to solve this problem by using RNNs, LSTMs, etc. This work examines the transformer models like BERT, RoBERTa, etc. We empirically demonstrate the effectiveness of transformer models and data augmentation for automated essay grading across many topics using a single model.
comment: Accepted at ICCMST 2022
♻ ☆ It's Just a Matter of Time: Detecting Depression with Time-Enriched Multimodal Transformers ECIR 2023
Depression detection from user-generated content on the internet has been a long-lasting topic of interest in the research community, providing valuable screening tools for psychologists. The ubiquitous use of social media platforms lays out the perfect avenue for exploring mental health manifestations in posts and interactions with other users. Current methods for depression detection from social media mainly focus on text processing, and only a few also utilize images posted by users. In this work, we propose a flexible time-enriched multimodal transformer architecture for detecting depression from social media posts, using pretrained models for extracting image and text embeddings. Our model operates directly at the user-level, and we enrich it with the relative time between posts by using time2vec positional embeddings. Moreover, we propose another model variant, which can operate on randomly sampled and unordered sets of posts to be more robust to dataset noise. We show that our method, using EmoBERTa and CLIP embeddings, surpasses other methods on two multimodal datasets, obtaining state-of-the-art results of 0.931 F1 score on a popular multimodal Twitter dataset, and 0.902 F1 score on the only multimodal Reddit dataset.
comment: Accepted at ECIR 2023
♻ ☆ Translate First Reorder Later: Leveraging Monotonicity in Semantic Parsing ACL
Prior work in semantic parsing has shown that conventional seq2seq models fail at compositional generalization tasks. This limitation led to a resurgence of methods that model alignments between sentences and their corresponding meaning representations, either implicitly through latent variables or explicitly by taking advantage of alignment annotations. We take the second direction and propose TPOL, a two-step approach that first translates input sentences monotonically and then reorders them to obtain the correct output. This is achieved with a modular framework comprising a Translator and a Reorderer component. We test our approach on two popular semantic parsing datasets. Our experiments show that by means of the monotonic translations, TPOL can learn reliable lexico-logical patterns from aligned data, significantly improving compositional generalization both over conventional seq2seq models, as well as over other approaches that exploit gold alignments.
comment: Accepted at Findings of ACL: EACL 2023. 8 pages, 4 figures, 4 tables
♻ ☆ Causal-Discovery Performance of ChatGPT in the context of Neuropathic Pain Diagnosis
ChatGPT has demonstrated exceptional proficiency in natural language conversation, e.g., it can answer a wide range of questions while no previous large language models can. Thus, we would like to push its limit and explore its ability to answer causal discovery questions by using a medical benchmark (Tu et al. 2019) in causal discovery.
♻ ☆ Show me your NFT and I tell you how it will perform: Multimodal representation learning for NFT selling price prediction
Non-Fungible Tokens (NFTs) represent deeds of ownership, based on blockchain technologies and smart contracts, of unique crypto assets on digital art forms (e.g., artworks or collectibles). In the spotlight after skyrocketing in 2021, NFTs have attracted the attention of crypto enthusiasts and investors intent on placing promising investments in this profitable market. However, the NFT financial performance prediction has not been widely explored to date. In this work, we address the above problem based on the hypothesis that NFT images and their textual descriptions are essential proxies to predict the NFT selling prices. To this purpose, we propose MERLIN, a novel multimodal deep learning framework designed to train Transformer-based language and visual models, along with graph neural network models, on collections of NFTs' images and texts. A key aspect in MERLIN is its independence on financial features, as it exploits only the primary data a user interested in NFT trading would like to deal with, i.e., NFT images and textual descriptions. By learning dense representations of such data, a price-category classification task is performed by MERLIN models, which can also be tuned according to user preferences in the inference phase to mimic different risk-return investment profiles. Experimental evaluation on a publicly available dataset has shown that MERLIN models achieve significant performances according to several financial assessment criteria, fostering profitable investments, and also beating baseline machine-learning classifiers based on financial features.
comment: Accepted paper at The ACM Web Conference 2023, April 30--May 04, 2023, Austin, Texas, USA
♻ ☆ StockEmotions: Discover Investor Emotions for Financial Sentiment Analysis and Multivariate Time Series AAAI-23
There has been growing interest in applying NLP techniques in the financial domain, however, resources are extremely limited. This paper introduces StockEmotions, a new dataset for detecting emotions in the stock market that consists of 10,000 English comments collected from StockTwits, a financial social media platform. Inspired by behavioral finance, it proposes 12 fine-grained emotion classes that span the roller coaster of investor emotion. Unlike existing financial sentiment datasets, StockEmotions presents granular features such as investor sentiment classes, fine-grained emotions, emojis, and time series data. To demonstrate the usability of the dataset, we perform a dataset analysis and conduct experimental downstream tasks. For financial sentiment/emotion classification tasks, DistilBERT outperforms other baselines, and for multivariate time series forecasting, a Temporal Attention LSTM model combining price index, text, and emotion features achieves the best performance than using a single feature.
comment: Preprint - Accepted by the AAAI-23 Bridge Program (AI for Financial Services)
♻ ☆ Joint Reasoning on Hybrid-knowledge sources for Task-Oriented Dialog
Traditional systems designed for task oriented dialog utilize knowledge present only in structured knowledge sources to generate responses. However, relevant information required to generate responses may also reside in unstructured sources, such as documents. Recent state of the art models such as HyKnow and SeKnow aimed at overcoming these challenges make limiting assumptions about the knowledge sources. For instance, these systems assume that certain types of information, such as a phone number, is always present in a structured knowledge base (KB) while information about aspects such as entrance ticket prices, would always be available in documents. In this paper, we create a modified version of the MutliWOZ-based dataset prepared by SeKnow to demonstrate how current methods have significant degradation in performance when strict assumptions about the source of information are removed. Then, in line with recent work exploiting pre-trained language models, we fine-tune a BART based model using prompts for the tasks of querying knowledge sources, as well as, for response generation, without making assumptions about the information present in each knowledge source. Through a series of experiments, we demonstrate that our model is robust to perturbations to knowledge modality (source of information), and that it can fuse information from structured as well as unstructured knowledge to generate responses.
♻ ☆ Editing Language Model-based Knowledge Graph Embeddings
Recently decades have witnessed the empirical success of framing Knowledge Graph (KG) embeddings via language models. However, language model-based KG embeddings are usually deployed as static artifacts, which are challenging to modify without re-training after deployment. To address this issue, we propose a new task of editing language model-based KG embeddings in this paper. The proposed task aims to enable data-efficient and fast updates to KG embeddings without damaging the performance of the rest. We build four new datasets: E-FB15k237, A-FB15k237, E-WN18RR, and A-WN18RR, and evaluate several knowledge editing baselines demonstrating the limited ability of previous models to handle the proposed challenging task. We further propose a simple yet strong baseline dubbed KGEditor, which utilizes additional parametric layers of the hyper network to edit/add facts. Comprehensive experimental results demonstrate that KGEditor can perform better when updating specific facts while not affecting the rest with low training resources. Code and datasets will be available in https://github.com/zjunlp/PromptKG/tree/main/deltaKG.
comment: Work in progress and the project website is https://zjunlp.github.io/project/KGE_Editing/
♻ ☆ LIQUID: A Framework for List Question Answering Dataset Generation AAAI 2023
Question answering (QA) models often rely on large-scale training datasets, which necessitates the development of a data generation framework to reduce the cost of manual annotations. Although several recent studies have aimed to generate synthetic questions with single-span answers, no study has been conducted on the creation of list questions with multiple, non-contiguous spans as answers. To address this gap, we propose LIQUID, an automated framework for generating list QA datasets from unlabeled corpora. We first convert a passage from Wikipedia or PubMed into a summary and extract named entities from the summarized text as candidate answers. This allows us to select answers that are semantically correlated in context and is, therefore, suitable for constructing list questions. We then create questions using an off-the-shelf question generator with the extracted entities and original passage. Finally, iterative filtering and answer expansion are performed to ensure the accuracy and completeness of the answers. Using our synthetic data, we significantly improve the performance of the previous best list QA models by exact-match F1 scores of 5.0 on MultiSpanQA, 1.9 on Quoref, and 2.8 averaged across three BioASQ benchmarks.
comment: AAAI 2023
♻ ☆ Transformers as Algorithms: Generalization and Stability in In-context Learning
In-context learning (ICL) is a type of prompting where a transformer model operates on a sequence of (input, output) examples and performs inference on-the-fly. In this work, we formalize in-context learning as an algorithm learning problem where a transformer model implicitly constructs a hypothesis function at inference-time. We first explore the statistical aspects of this abstraction through the lens of multitask learning: We obtain generalization bounds for ICL when the input prompt is (1) a sequence of i.i.d. (input, label) pairs or (2) a trajectory arising from a dynamical system. The crux of our analysis is relating the excess risk to the stability of the algorithm implemented by the transformer. We characterize when transformer/attention architecture provably obeys the stability condition and also provide empirical verification. For generalization on unseen tasks, we identify an inductive bias phenomenon in which the transfer learning risk is governed by the task complexity and the number of MTL tasks in a highly predictable manner. Finally, we provide numerical evaluations that (1) demonstrate transformers can indeed implement near-optimal algorithms on classical regression problems with i.i.d. and dynamic data, (2) provide insights on stability, and (3) verify our theoretical predictions.
comment: Revised version significantly improves the stability guarantees and provides new experiments
♻ ☆ On Robustness of Prompt-based Semantic Parsing with Large Pre-trained Language Model: An Empirical Study on Codex EACL2023
Semantic parsing is a technique aimed at constructing a structured representation of the meaning of a natural-language question. Recent advancements in few-shot language models trained on code have demonstrated superior performance in generating these representations compared to traditional unimodal language models, which are trained on downstream tasks. Despite these advancements, existing fine-tuned neural semantic parsers are susceptible to adversarial attacks on natural-language inputs. While it has been established that the robustness of smaller semantic parsers can be enhanced through adversarial training, this approach is not feasible for large language models in real-world scenarios, as it requires both substantial computational resources and expensive human annotation on in-domain semantic parsing data. This paper presents the first empirical study on the adversarial robustness of a large prompt-based language model of code, \codex. Our results demonstrate that the state-of-the-art (SOTA) code-language models are vulnerable to carefully crafted adversarial examples. To address this challenge, we propose methods for improving robustness without the need for significant amounts of labeled data or heavy computational resources.
comment: Accepted at EACL2023 (main)
♻ ☆ Construction of English Resume Corpus and Test with Pre-trained Language Models
Information extraction(IE) has always been one of the essential tasks of NLP. Moreover, one of the most critical application scenarios of information extraction is the information extraction of resumes. Constructed text is obtained by classifying each part of the resume. It is convenient to store these texts for later search and analysis. Furthermore, the constructed resume data can also be used in the AI resume screening system. Significantly reduce the labor cost of HR. This study aims to transform the information extraction task of resumes into a simple sentence classification task. Based on the English resume dataset produced by the prior study. The classification rules are improved to create a larger and more fine-grained classification dataset of resumes. This corpus is also used to test some current mainstream Pre-training language models (PLMs) performance.Furthermore, in order to explore the relationship between the number of training samples and the correctness rate of the resume dataset, we also performed comparison experiments with training sets of different train set sizes.The final multiple experimental results show that the resume dataset with improved annotation rules and increased sample size of the dataset improves the accuracy of the original resume dataset.
comment: The Association for Natural Language Processing in Japan. Submit in NLP2023
♻ ☆ Bidirectional Language Models Are Also Few-shot Learners ICLR 2023
Large language models such as GPT-3 (Brown et al., 2020) can perform arbitrary tasks without undergoing fine-tuning after being prompted with only a few labeled examples. An arbitrary task can be reformulated as a natural language prompt, and a language model can be asked to generate the completion, indirectly performing the task in a paradigm known as prompt-based learning. To date, emergent prompt-based learning capabilities have mainly been demonstrated for unidirectional language models. However, bidirectional language models pre-trained on denoising objectives such as masked language modeling produce stronger learned representations for transfer learning. This motivates the possibility of prompting bidirectional models, but their pre-training objectives have made them largely incompatible with the existing prompting paradigm. We present SAP (Sequential Autoregressive Prompting), a technique that enables the prompting of bidirectional models. Utilizing the machine translation task as a case study, we prompt the bidirectional mT5 model (Xue et al., 2021) with SAP and demonstrate its few-shot and zero-shot translations outperform the few-shot translations of unidirectional models like GPT-3 and XGLM (Lin et al., 2021), despite mT5's approximately 50% fewer parameters. We further show SAP is effective on question answering and summarization. For the first time, our results demonstrate prompt-based learning is an emergent property of a broader class of language models, rather than only unidirectional models.
comment: To appear at ICLR 2023
♻ ☆ Knowledge Distillation $\approx$ Label Smoothing: Fact or Fallacy?
Contrary to its original interpretation as a facilitator of knowledge transfer from one model to another, some recent studies have suggested that knowledge distillation (KD) is instead a form of regularization. Perhaps the strongest support of all for this claim is found in its apparent similarities with label smoothing (LS). This paper investigates the stated equivalence of these two methods by examining the predictive uncertainties of the models they train. Experiments on four text classification tasks involving teachers and students of different capacities show that: (a) In most settings, KD and LS drive model uncertainty (entropy) in completely opposite directions, and (b) In KD, the student's predictive uncertainty is a direct function of that of its teacher, reinforcing the knowledge transfer view.
♻ ☆ BiFSMNv2: Pushing Binary Neural Networks for Keyword Spotting to Real-Network Performance
Deep neural networks, such as the Deep-FSMN, have been widely studied for keyword spotting (KWS) applications while suffering expensive computation and storage. Therefore, network compression technologies like binarization are studied to deploy KWS models on edge. In this paper, we present a strong yet efficient binary neural network for KWS, namely BiFSMNv2, pushing it to the real-network accuracy performance. First, we present a Dual-scale Thinnable 1-bit-Architecture to recover the representation capability of the binarized computation units by dual-scale activation binarization and liberate the speedup potential from an overall architecture perspective. Second, we also construct a Frequency Independent Distillation scheme for KWS binarization-aware training, which distills the high and low-frequency components independently to mitigate the information mismatch between full-precision and binarized representations. Moreover, we propose the Learning Propagation Binarizer, a general and efficient binarizer that enables the forward and backward propagation of binary KWS networks to be continuously improved through learning. We implement and deploy the BiFSMNv2 on ARMv8 real-world hardware with a novel Fast Bitwise Computation Kernel, which is proposed to fully utilize registers and increase instruction throughput. Comprehensive experiments show our BiFSMNv2 outperforms existing binary networks for KWS by convincing margins across different datasets and achieves comparable accuracy with the full-precision networks (only a tiny 1.51% drop on Speech Commands V1-12). We highlight that benefiting from the compact architecture and optimized hardware kernel, BiFSMNv2 can achieve an impressive 25.1x speedup and 20.2x storage-saving on edge hardware.
comment: arXiv admin note: text overlap with arXiv:2202.06483
♻ ☆ Morality, Machines and the Interpretation Problem: A Value-based, Wittgensteinian Approach to Building Moral Agents
We present what we call the Interpretation Problem, whereby any rule in symbolic form is open to infinite interpretation in ways that we might disapprove of and argue that any attempt to build morality into machines is subject to it. We show how the Interpretation Problem in Artificial Intelligence is an illustration of Wittgenstein's general claim that no rule can contain the criteria for its own application, and that the risks created by this problem escalate in proportion to the degree to which to machine is causally connected to the world, in what we call the Law of Interpretative Exposure. Using game theory, we attempt to define the structure of normative spaces and argue that any rule-following within a normative space is guided by values that are external to that space and which cannot themselves be represented as rules. In light of this, we categorise the types of mistakes an artificial moral agent could make into Mistakes of Intention and Instrumental Mistakes, and we propose ways of building morality into machines by getting them to interpret the rules we give in accordance with these external values, through explicit moral reasoning, the Show, not Tell paradigm, the adjustment of causal power and structure of the agent, and relational values, with the ultimate aim that the machine develop a virtuous character and that the impact of the Interpretation Problem is minimised.
Computer Vision and Pattern Recognition 88
☆ Zero-shot Image-to-Image Translation
Large-scale text-to-image generative models have shown their remarkable ability to synthesize diverse and high-quality images. However, it is still challenging to directly apply these models for editing real images for two reasons. First, it is hard for users to come up with a perfect text prompt that accurately describes every visual detail in the input image. Second, while existing models can introduce desirable changes in certain regions, they often dramatically alter the input content and introduce unexpected changes in unwanted regions. In this work, we propose pix2pix-zero, an image-to-image translation method that can preserve the content of the original image without manual prompting. We first automatically discover editing directions that reflect desired edits in the text embedding space. To preserve the general content structure after editing, we further propose cross-attention guidance, which aims to retain the cross-attention maps of the input image throughout the diffusion process. In addition, our method does not need additional training for these edits and can directly use the existing pre-trained text-to-image diffusion model. We conduct extensive experiments and show that our method outperforms existing and concurrent works for both real and synthetic image editing.
comment: website: https://pix2pixzero.github.io/
☆ AIM: Adapting Image Models for Efficient Video Action Recognition ICLR 2023
Recent vision transformer based video models mostly follow the ``image pre-training then finetuning" paradigm and have achieved great success on multiple video benchmarks. However, full finetuning such a video model could be computationally expensive and unnecessary, given the pre-trained image transformer models have demonstrated exceptional transferability. In this work, we propose a novel method to Adapt pre-trained Image Models (AIM) for efficient video understanding. By freezing the pre-trained image model and adding a few lightweight Adapters, we introduce spatial adaptation, temporal adaptation and joint adaptation to gradually equip an image model with spatiotemporal reasoning capability. We show that our proposed AIM can achieve competitive or even better performance than prior arts with substantially fewer tunable parameters on four video action recognition benchmarks. Thanks to its simplicity, our method is also generally applicable to different image pre-trained models, which has the potential to leverage more powerful image foundation models in the future. The project webpage is \url{https://adapt-image-models.github.io/}.
comment: Accepted to ICLR 2023. Project webpage is at https://adapt-image-models.github.io/
☆ V1T: large-scale mouse V1 response prediction using a Vision Transformer
Accurate predictive models of the visual cortex neural response to natural visual stimuli remain a challenge in computational neuroscience. In this work, we introduce V1T, a novel Vision Transformer based architecture that learns a shared visual and behavioral representation across animals. We evaluate our model on two large datasets recorded from mouse primary visual cortex and outperform previous convolution-based models by more than 12.7% in prediction performance. Moreover, we show that the attention weights learned by the Transformer correlate with the population receptive fields. Our model thus sets a new benchmark for neural response prediction and captures characteristic features of the visual cortex.
☆ SurgT: Soft-Tissue Tracking for Robotic Surgery, Benchmark and Challenge
This paper introduces the SurgT MICCAI 2022 challenge and its first results. There were two purposes for the creation of this challenge: (1) the establishment of the first standardised benchmark for the research community to assess soft-tissue trackers; and (2) to encourage the development of unsupervised deep learning methods, given the lack of annotated data in surgery. A dataset of 157 stereo endoscopic videos from 20 clinical cases, along with stereo camera calibration parameters, are provided. The participants were tasked with the development of algorithms to track a bounding box on each stereo endoscopic video. At the end of the challenge, the developed methods were assessed on a previously hidden test subset. This assessment uses benchmarking metrics that were purposely developed for this challenge and are now available online. The teams were ranked according to their Expected Average Overlap (EAO) score, which is a weighted average of Intersection over Union (IoU) scores. The top team achieved an EAO score of 0.583 in the test subset. Tracking soft-tissue using unsupervised algorithms was found to be achievable. The dataset and benchmarking tool have been successfully created and made publicly available online. This challenge is expected to contribute to the development of autonomous robotic surgery, and other digital surgical technologies.
☆ RLSbench: Domain Adaptation Under Relaxed Label Shift
Despite the emergence of principled methods for domain adaptation under label shift, the sensitivity of these methods for minor shifts in the class conditional distributions remains precariously under explored. Meanwhile, popular deep domain adaptation heuristics tend to falter when faced with shifts in label proportions. While several papers attempt to adapt these heuristics to accommodate shifts in label proportions, inconsistencies in evaluation criteria, datasets, and baselines, make it hard to assess the state of the art. In this paper, we introduce RLSbench, a large-scale relaxed label shift benchmark, consisting of >500 distribution shift pairs that draw on 14 datasets across vision, tabular, and language modalities and compose them with varying label proportions. First, we evaluate 13 popular domain adaptation methods, demonstrating more widespread failures under label proportion shifts than were previously known. Next, we develop an effective two-step meta-algorithm that is compatible with most deep domain adaptation heuristics: (i) pseudo-balance the data at each epoch; and (ii) adjust the final classifier with (an estimate of) target label distribution. The meta-algorithm improves existing domain adaptation heuristics often by 2--10\% accuracy points under extreme label proportion shifts and has little (i.e., <0.5\%) effect when label proportions do not shift. We hope that these findings and the availability of RLSbench will encourage researchers to rigorously evaluate proposed methods in relaxed label shift settings. Code is publicly available at https://github.com/acmi-lab/RLSbench.
☆ DDM$^2$: Self-Supervised Diffusion MRI Denoising with Generative Diffusion Models ICLR 2023
Magnetic resonance imaging (MRI) is a common and life-saving medical imaging technique. However, acquiring high signal-to-noise ratio MRI scans requires long scan times, resulting in increased costs and patient discomfort, and decreased throughput. Thus, there is great interest in denoising MRI scans, especially for the subtype of diffusion MRI scans that are severely SNR-limited. While most prior MRI denoising methods are supervised in nature, acquiring supervised training datasets for the multitude of anatomies, MRI scanners, and scan parameters proves impractical. Here, we propose Denoising Diffusion Models for Denoising Diffusion MRI (DDM$^2$), a self-supervised denoising method for MRI denoising using diffusion denoising generative models. Our three-stage framework integrates statistic-based denoising theory into diffusion models and performs denoising through conditional generation. During inference, we represent input noisy measurements as a sample from an intermediate posterior distribution within the diffusion Markov chain. We conduct experiments on 4 real-world in-vivo diffusion MRI datasets and show that our DDM$^2$ demonstrates superior denoising performances ascertained with clinically-relevant visual qualitative and quantitative metrics.
comment: To appear in ICLR 2023
☆ Detection and Localization of Melanoma Skin Cancer in Histopathological Whole Slide Images
Melanoma diagnosed and treated in its early stages can increase the survival rate. A projected increase in skin cancer incidents and a dearth of dermatopathologists have emphasized the need for computational pathology (CPATH) systems. CPATH systems with deep learning (DL) models have the potential to identify the presence of melanoma by exploiting underlying morphological and cellular features. This paper proposes a DL method to detect melanoma and distinguish between normal skin and benign/malignant melanocytic lesions in Whole Slide Images (WSI). Our method detects lesions with high accuracy and localizes them on a WSI to identify potential regions of interest for pathologists. Interestingly, our DL method relies on using a single CNN network to create localization maps first and use them to perform slide-level predictions to determine patients who have melanoma. Our best model provides favorable patch-wise classification results with a 0.992 F1 score and 0.99 sensitivity on unseen data.
comment: Submitted to EUSIPCO 23
☆ Structure and Content-Guided Video Synthesis with Diffusion Models
Text-guided generative diffusion models unlock powerful image creation and editing tools. While these have been extended to video generation, current approaches that edit the content of existing footage while retaining structure require expensive re-training for every input or rely on error-prone propagation of image edits across frames. In this work, we present a structure and content-guided video diffusion model that edits videos based on visual or textual descriptions of the desired output. Conflicts between user-provided content edits and structure representations occur due to insufficient disentanglement between the two aspects. As a solution, we show that training on monocular depth estimates with varying levels of detail provides control over structure and content fidelity. Our model is trained jointly on images and videos which also exposes explicit control of temporal consistency through a novel guidance method. Our experiments demonstrate a wide variety of successes; fine-grained control over output characteristics, customization based on a few reference images, and a strong user preference towards results by our model.
comment: Project page at https://research.runwayml.com/gen1
☆ OTRE: Where Optimal Transport Guided Unpaired Image-to-Image Translation Meets Regularization by Enhancing
Non-mydriatic retinal color fundus photography (CFP) is widely available due to the advantage of not requiring pupillary dilation, however, is prone to poor quality due to operators, systemic imperfections, or patient-related causes. Optimal retinal image quality is mandated for accurate medical diagnoses and automated analyses. Herein, we leveraged the \emph{Optimal Transport (OT)} theory to propose an unpaired image-to-image translation scheme for mapping low-quality retinal CFPs to high-quality counterparts. Furthermore, to improve the flexibility, robustness, and applicability of our image enhancement pipeline in the clinical practice, we generalized a state-of-the-art model-based image reconstruction method, regularization by denoising, by plugging in priors learned by our OT-guided image-to-image translation network. We named it as \emph{regularization by enhancing (RE)}. We validated the integrated framework, OTRE, on three publicly available retinal image datasets by assessing the quality after enhancement and their performance on various downstream tasks, including diabetic retinopathy grading, vessel segmentation, and diabetic lesion segmentation. The experimental results demonstrated the superiority of our proposed framework over some state-of-the-art unsupervised competitors and a state-of-the-art supervised method.
comment: Accepted as a conference paper to The 28th biennial international conference on Information Processing in Medical Imaging (IPMI 2023)
☆ Neural Collapse Inspired Feature-Classifier Alignment for Few-Shot Class Incremental Learning ICLR 2023
Few-shot class-incremental learning (FSCIL) has been a challenging problem as only a few training samples are accessible for each novel class in the new sessions. Finetuning the backbone or adjusting the classifier prototypes trained in the prior sessions would inevitably cause a misalignment between the feature and classifier of old classes, which explains the well-known catastrophic forgetting problem. In this paper, we deal with this misalignment dilemma in FSCIL inspired by the recently discovered phenomenon named neural collapse, which reveals that the last-layer features of the same class will collapse into a vertex, and the vertices of all classes are aligned with the classifier prototypes, which are formed as a simplex equiangular tight frame (ETF). It corresponds to an optimal geometric structure for classification due to the maximized Fisher Discriminant Ratio. We propose a neural collapse inspired framework for FSCIL. A group of classifier prototypes are pre-assigned as a simplex ETF for the whole label space, including the base session and all the incremental sessions. During training, the classifier prototypes are not learnable, and we adopt a novel loss function that drives the features into their corresponding prototypes. Theoretical analysis shows that our method holds the neural collapse optimality and does not break the feature-classifier alignment in an incremental fashion. Experiments on the miniImageNet, CUB-200, and CIFAR-100 datasets demonstrate that our proposed framework outperforms the state-of-the-art performances. Code address: https://github.com/NeuralCollapseApplications/FSCIL
comment: ICLR 2023 (Notable-top-25%)
☆ Optimal Transport Guided Unsupervised Learning for Enhancing low-quality Retinal Images
Real-world non-mydriatic retinal fundus photography is prone to artifacts, imperfections and low-quality when certain ocular or systemic co-morbidities exist. Artifacts may result in inaccuracy or ambiguity in clinical diagnoses. In this paper, we proposed a simple but effective end-to-end framework for enhancing poor-quality retinal fundus images. Leveraging the optimal transport theory, we proposed an unpaired image-to-image translation scheme for transporting low-quality images to their high-quality counterparts. We theoretically proved that a Generative Adversarial Networks (GAN) model with a generator and discriminator is sufficient for this task. Furthermore, to mitigate the inconsistency of information between the low-quality images and their enhancements, an information consistency mechanism was proposed to maximally maintain structural consistency (optical discs, blood vessels, lesions) between the source and enhanced domains. Extensive experiments were conducted on the EyeQ dataset to demonstrate the superiority of our proposed method perceptually and quantitatively.
comment: Accepted as a conference paper to 20th IEEE International Symposium on Biomedical Imaging(ISBI 2023)
☆ Learning disentangled representations for explainable chest X-ray classification using Dirichlet VAEs SP
This study explores the use of the Dirichlet Variational Autoencoder (DirVAE) for learning disentangled latent representations of chest X-ray (CXR) images. Our working hypothesis is that distributional sparsity, as facilitated by the Dirichlet prior, will encourage disentangled feature learning for the complex task of multi-label classification of CXR images. The DirVAE is trained using CXR images from the CheXpert database, and the predictive capacity of multi-modal latent representations learned by DirVAE models is investigated through implementation of an auxiliary multi-label classification task, with a view to enforce separation of latent factors according to class-specific features. The predictive performance and explainability of the latent space learned using the DirVAE were quantitatively and qualitatively assessed, respectively, and compared with a standard Gaussian prior-VAE (GVAE). We introduce a new approach for explainable multi-label classification in which we conduct gradient-guided latent traversals for each class of interest. Study findings indicate that the DirVAE is able to disentangle latent factors into class-specific visual features, a property not afforded by the GVAE, and achieve a marginal increase in predictive performance relative to GVAE. We generate visual examples to show that our explainability method, when applied to the trained DirVAE, is able to highlight regions in CXR images that are clinically relevant to the class(es) of interest and additionally, can identify cases where classification relies on spurious feature correlations.
comment: 13 pages, 8 figures, to be published in SPIE Medical Imaging 2023
☆ ConvoWaste: An Automatic Waste Segregation Machine Using Deep Learning
Nowadays, proper urban waste management is one of the biggest concerns for maintaining a green and clean environment. An automatic waste segregation system can be a viable solution to improve the sustainability of the country and boost the circular economy. This paper proposes a machine to segregate waste into different parts with the help of a smart object detection algorithm using ConvoWaste in the field of deep convolutional neural networks (DCNN) and image processing techniques. In this paper, deep learning and image processing techniques are applied to precisely classify the waste, and the detected waste is placed inside the corresponding bins with the help of a servo motor-based system. This machine has the provision to notify the responsible authority regarding the waste level of the bins and the time to trash out the bins filled with garbage by using the ultrasonic sensors placed in each bin and the dual-band GSM-based communication technology. The entire system is controlled remotely through an Android app in order to dump the separated waste in the desired place thanks to its automation properties. The use of this system can aid in the process of recycling resources that were initially destined to become waste, utilizing natural resources, and turning these resources back into usable products. Thus, the system helps fulfill the criteria of a circular economy through resource optimization and extraction. Finally, the system is designed to provide services at a low cost while maintaining a high level of accuracy in terms of technological advancement in the field of artificial intelligence (AI). We have gotten 98% accuracy for our ConvoWaste deep learning model.
comment: Accepted at ICREST 2023
☆ Integrating Eye-Gaze Data into CXR DL Approaches: A Preliminary study
This paper proposes a novel multimodal DL architecture incorporating medical images and eye-tracking data for abnormality detection in chest x-rays. Our results show that applying eye gaze data directly into DL architectures does not show superior predictive performance in abnormality detection chest X-rays. These results support other works in the literature and suggest that human-generated data, such as eye gaze, needs a more thorough investigation before being applied to DL architectures.
comment: A version of this paper has been accepted for presentation at the 2nd XR Health workshop - XR Technologies for Healthcare and Wellbeing https://ieeevr.org/2023/contribute/workshoppapers/#XRHealth
☆ Private GANs, Revisited
We show that the canonical approach for training differentially private GANs -- updating the discriminator with differentially private stochastic gradient descent (DPSGD) -- can yield significantly improved results after modifications to training. Existing instantiations of this approach neglect to consider how adding noise only to discriminator updates disrupts the careful balance between the generator and discriminator necessary for successful GAN training. We show that a simple fix -- taking more discriminator steps between generator steps -- restores parity and improves results. Additionally, with the goal of restoring parity between the generator and discriminator, we experiment with other modifications to improve discriminator training and see further improvements in generation quality. Our results demonstrate that on standard benchmarks, DPSGD outperforms all alternative GAN privatization schemes.
☆ Generating Evidential BEV Maps in Continuous Driving Space
Safety is critical for autonomous driving, and one aspect of improving safety is to accurately capture the uncertainties of the perception system, especially knowing the unknown. Different from only providing deterministic or probabilistic results, e.g., probabilistic object detection, that only provide partial information for the perception scenario, we propose a complete probabilistic model named GevBEV. It interprets the 2D driving space as a probabilistic Bird's Eye View (BEV) map with point-based spatial Gaussian distributions, from which one can draw evidence as the parameters for the categorical Dirichlet distribution of any new sample point in the continuous driving space. The experimental results show that GevBEV not only provides more reliable uncertainty quantification but also outperforms the previous works on the benchmark OPV2V of BEV map interpretation for cooperative perception. A critical factor in cooperative perception is the data transmission size through the communication channels. GevBEV helps reduce communication overhead by selecting only the most important information to share from the learned uncertainty, reducing the average information communicated by 80% with a slight performance drop.
☆ LexLIP: Lexicon-Bottlenecked Language-Image Pre-Training for Large-Scale Image-Text Retrieval
Image-text retrieval (ITR) is a task to retrieve the relevant images/texts, given the query from another modality. The conventional dense retrieval paradigm relies on encoding images and texts into dense representations using dual-stream encoders, however, it faces challenges with low retrieval speed in large-scale retrieval scenarios. In this work, we propose the lexicon-weighting paradigm, where sparse representations in vocabulary space are learned for images and texts to take advantage of the bag-of-words models and efficient inverted indexes, resulting in significantly reduced retrieval latency. A crucial gap arises from the continuous nature of image data, and the requirement for a sparse vocabulary space representation. To bridge this gap, we introduce a novel pre-training framework, Lexicon-Bottlenecked Language-Image Pre-Training (LexLIP), that learns importance-aware lexicon representations. This framework features lexicon-bottlenecked modules between the dual-stream encoders and weakened text decoders, allowing for constructing continuous bag-of-words bottlenecks to learn lexicon-importance distributions. Upon pre-training with same-scale data, our LexLIP achieves state-of-the-art performance on two benchmark ITR datasets, MSCOCO and Flickr30k. Furthermore, in large-scale retrieval scenarios, LexLIP outperforms CLIP with a 5.5 ~ 221.3X faster retrieval speed and 13.2 ~ 48.8X less index storage memory.
☆ GAT: Guided Adversarial Training with Pareto-optimal Auxiliary Tasks
While leveraging additional training data is well established to improve adversarial robustness, it incurs the unavoidable cost of data collection and the heavy computation to train models. To mitigate the costs, we propose \textit{Guided Adversarial Training } (GAT), a novel adversarial training technique that exploits auxiliary tasks under a limited set of training data. Our approach extends single-task models into multi-task models during the min-max optimization of adversarial training, and drives the loss optimization with a regularization of the gradient curvature across multiple tasks. GAT leverages two types of auxiliary tasks: self-supervised tasks, where the labels are generated automatically, and domain-knowledge tasks, where human experts provide additional labels. Experimentally, under limited data, GAT increases the robust accuracy on CIFAR-10 up to four times (from 11% to 42% robust accuracy) and the robust AUC of CheXpert medical imaging dataset from 50\% to 83\%. On the full CIFAR-10 dataset, GAT outperforms eight state-of-the-art adversarial training strategies. Our large study across five datasets and six tasks demonstrates that task augmentation is an efficient alternative to data augmentation, and can be key to achieving both clean and robust performances.
☆ Neural Document Unwarping using Coupled Grids
Restoring the original, flat appearance of a printed document from casual photographs of bent and wrinkled pages is a common everyday problem. In this paper we propose a novel method for grid-based single-image document unwarping. Our method performs geometric distortion correction via a deep fully convolutional neural network that learns to predict the 3D grid mesh of the document and the corresponding 2D unwarping grid in a multi-task fashion, implicitly encoding the coupling between the shape of a 3D object and its 2D image. We additionally create and publish our own dataset, called UVDoc, which combines pseudo-photorealistic document images with ground truth grid-based physical 3D and unwarping information, allowing unwarping models to train on data that is more realistic in appearance than the commonly used synthetic Doc3D dataset, whilst also being more physically accurate. Our dataset is labeled with all the information necessary to train our unwarping network, without having to engineer separate loss functions that can deal with the lack of ground-truth typically found in document in the wild datasets. We include a thorough evaluation that demonstrates that our dual-task unwarping network trained on a mix of synthetic and pseudo-photorealistic images achieves state-of-the-art performance on the DocUNet benchmark dataset. Our code, results and UVDoc dataset will be made publicly available upon publication.
☆ Intra-operative Brain Tumor Detection with Deep Learning-Optimized Hyperspectral Imaging SP
Surgery for gliomas (intrinsic brain tumors), especially when low-grade, is challenging due to the infiltrative nature of the lesion. Currently, no real-time, intra-operative, label-free and wide-field tool is available to assist and guide the surgeon to find the relevant demarcations for these tumors. While marker-based methods exist for the high-grade glioma case, there is no convenient solution available for the low-grade case; thus, marker-free optical techniques represent an attractive option. Although RGB imaging is a standard tool in surgical microscopes, it does not contain sufficient information for tissue differentiation. We leverage the richer information from hyperspectral imaging (HSI), acquired with a snapscan camera in the 468-787 nm range, coupled to a surgical microscope, to build a deep-learning-based diagnostic tool for cancer resection with potential for intra-operative guidance. However, the main limitation of the HSI snapscan camera is the image acquisition time, limiting its widespread deployment in the operation theater. Here, we investigate the effect of HSI channel reduction and pre-selection to scope the design space for the development of cheaper and faster sensors. Neural networks are used to identify the most important spectral channels for tumor tissue differentiation, optimizing the trade-off between the number of channels and precision to enable real-time intra-surgical application. We evaluate the performance of our method on a clinical dataset that was acquired during surgery on five patients. By demonstrating the possibility to efficiently detect low-grade glioma, these results can lead to better cancer resection demarcations, potentially improving treatment effectiveness and patient outcome.
comment: SPIE Photonics West 2023 conference Optical Biopsy XXI: Toward Real-Time Spectroscopic Imaging and Diagnosis. 18 pages, 11 figures
☆ Top-Down Beats Bottom-Up in 3D Instance Segmentation
Most 3D instance segmentation methods exploit a bottom-up strategy, typically including resource-exhaustive post-processing. For point grouping, bottom-up methods rely on prior assumptions about the objects in the form of hyperparameters, which are domain-specific and need to be carefully tuned. On the contrary, we address 3D instance segmentation with a TD3D: top-down, fully data-driven, simple approach trained in an end-to-end manner. With its straightforward fully-convolutional pipeline, it performs surprisingly well on the standard benchmarks: ScanNet v2, its extension ScanNet200, and S3DIS. Besides, our method is much faster on inference than the current state-of-the-art grouping-based approaches. Code is available at https://github.com/SamsungLabs/td3d .
☆ TR3D: Towards Real-Time Indoor 3D Object Detection
Recently, sparse 3D convolutions have changed 3D object detection. Performing on par with the voting-based approaches, 3D CNNs are memory-efficient and scale to large scenes better. However, there is still room for improvement. With a conscious, practice-oriented approach to problem-solving, we analyze the performance of such methods and localize the weaknesses. Applying modifications that resolve the found issues one by one, we end up with TR3D: a fast fully-convolutional 3D object detection model trained end-to-end, that achieves state-of-the-art results on the standard benchmarks, ScanNet v2, SUN RGB-D, and S3DIS. Moreover, to take advantage of both point cloud and RGB inputs, we introduce an early fusion of 2D and 3D features. We employ our fusion module to make conventional 3D object detection methods multimodal and demonstrate an impressive boost in performance. Our model with early feature fusion, which we refer to as TR3D+FF, outperforms existing 3D object detection approaches on the SUN RGB-D dataset. Overall, besides being accurate, both TR3D and TR3D+FF models are lightweight, memory-efficient, and fast, thereby marking another milestone on the way toward real-time 3D object detection. Code is available at https://github.com/SamsungLabs/tr3d .
☆ An Unsupervised Framework for Joint MRI Super Resolution and Gibbs Artifact Removal
The k-space data generated from magnetic resonance imaging (MRI) is only a finite sampling of underlying signals. Therefore, MRI images often suffer from low spatial resolution and Gibbs ringing artifacts. Previous studies tackled these two problems separately, where super resolution methods tend to enhance Gibbs artifacts, whereas Gibbs ringing removal methods tend to blur the images. It is also a challenge that high resolution ground truth is hard to obtain in clinical MRI. In this paper, we propose an unsupervised learning framework for both MRI super resolution and Gibbs artifacts removal without using high resolution ground truth. Furthermore, we propose regularization methods to improve the model's generalizability across out-of-distribution MRI images. We evaluated our proposed methods with other state-of-the-art methods on eight MRI datasets with various contrasts and anatomical structures. Our method not only achieves the best SR performance but also significantly reduces the Gibbs artifacts. Our method also demonstrates good generalizability across different datasets, which is beneficial to clinical applications where training data are usually scarce and biased.
comment: Accepted by the 28th biennial international conference on Information Processing in Medical Imaging (IPMI 2023)
☆ MixFormer: End-to-End Tracking with Iterative Mixed Attention CVPR 2022
Visual object tracking often employs a multi-stage pipeline of feature extraction, target information integration, and bounding box estimation. To simplify this pipeline and unify the process of feature extraction and target information integration, in this paper, we present a compact tracking framework, termed as MixFormer, built upon transformers. Our core design is to utilize the flexibility of attention operations, and propose a Mixed Attention Module (MAM) for simultaneous feature extraction and target information integration. This synchronous modeling scheme allows to extract target-specific discriminative features and perform extensive communication between target and search area. Based on MAM, we build our MixFormer trackers simply by stacking multiple MAMs and placing a localization head on top. Specifically, we instantiate two types of MixFormer trackers, a hierarchical tracker MixCvT, and a non-hierarchical tracker MixViT. For these two trackers, we investigate a series of pre-training methods and uncover the different behaviors between supervised pre-training and self-supervised pre-training in our MixFormer trackers. We also extend the masked pre-training to our MixFormer trackers and design the competitive TrackMAE pre-training technique. Finally, to handle multiple target templates during online tracking, we devise an asymmetric attention scheme in MAM to reduce computational cost, and propose an effective score prediction module to select high-quality templates. Our MixFormer trackers set a new state-of-the-art performance on seven tracking benchmarks, including LaSOT, TrackingNet, VOT2020, GOT-10k, OTB100 and UAV123. In particular, our MixViT-L achieves AUC score of 73.3% on LaSOT, 86.1% on TrackingNet, EAO of 0.584 on VOT2020, and AO of 75.7% on GOT-10k. Code and trained models will be made available at https://github.com/MCG-NJU/MixFormer.
comment: Extended version of the paper arXiv:2203.11082 presented at CVPR 2022. In particular, the extented MixViT-L achieves AUC score of 73.3% on LaSOT. Besides, we design a new TrackMAE pre-training method for tracking
☆ Stop overkilling simple tasks with black-box models and use transparent models instead
In recent years, the employment of deep learning methods has led to several significant breakthroughs in artificial intelligence. Different from traditional machine learning models, deep learning-based approaches are able to extract features autonomously from raw data. This allows for bypassing the feature engineering process, which is generally considered to be both error-prone and tedious. Moreover, deep learning strategies often outperform traditional models in terms of accuracy.
☆ Perception Datasets for Anomaly Detection in Autonomous Driving: A Survey
Deep neural networks (DNN) which are employed in perception systems for autonomous driving require a huge amount of data to train on, as they must reliably achieve high performance in all kinds of situations. However, these DNN are usually restricted to a closed set of semantic classes available in their training data, and are therefore unreliable when confronted with previously unseen instances. Thus, multiple perception datasets have been created for the evaluation of anomaly detection methods, which can be categorized into three groups: real anomalies in real-world, synthetic anomalies augmented into real-world and completely synthetic scenes. This survey provides a structured and, to the best of our knowledge, complete overview and comparison of perception datasets for anomaly detection in autonomous driving. Each chapter provides information about tasks and ground truth, context information, and licenses. Additionally, we discuss current weaknesses and gaps in existing datasets to underline the importance of developing further data.
☆ Fine-Grained Action Detection with RGB and Pose Information using Two Stream Convolutional Networks
As participants of the MediaEval 2022 Sport Task, we propose a two-stream network approach for the classification and detection of table tennis strokes. Each stream is a succession of 3D Convolutional Neural Network (CNN) blocks using attention mechanisms. Each stream processes different 4D inputs. Our method utilizes raw RGB data and pose information computed from MMPose toolbox. The pose information is treated as an image by applying the pose either on a black background or on the original RGB frame it has been computed from. Best performance is obtained by feeding raw RGB data to one stream, Pose + RGB (PRGB) information to the other stream and applying late fusion on the features. The approaches were evaluated on the provided TTStroke-21 data sets. We can report an improvement in stroke classification, reaching 87.3% of accuracy, while the detection does not outperform the baseline but still reaches an IoU of 0.349 and mAP of 0.110.
comment: Working note paper of the sport task of MediaEval 2022 in Bergen, Norway, 12-13 Jan 2023
☆ Baseline Method for the Sport Task of MediaEval 2022 with 3D CNNs using Attention Mechanisms
This paper presents the baseline method proposed for the Sports Video task part of the MediaEval 2022 benchmark. This task proposes two subtasks: stroke classification from trimmed videos, and stroke detection from untrimmed videos. This baseline addresses both subtasks. We propose two types of 3D-CNN architectures to solve the two subtasks. Both 3D-CNNs use Spatio-temporal convolutions and attention mechanisms. The architectures and the training process are tailored to solve the addressed subtask. This baseline method is shared publicly online to help the participants in their investigation and alleviate eventually some aspects of the task such as video processing, training method, evaluation and submission routine. The baseline method reaches 86.4% of accuracy with our v2 model for the classification subtask. For the detection subtask, the baseline reaches a mAP of 0.131 and IoU of 0.515 with our v1 model.
comment: Baseline paper for the sport Task of MediaEval 2022
☆ AMD-HookNet for Glacier Front Segmentation
Knowledge on changes in glacier calving front positions is important for assessing the status of glaciers. Remote sensing imagery provides the ideal database for monitoring calving front positions, however, it is not feasible to perform this task manually for all calving glaciers globally due to time-constraints. Deep learning-based methods have shown great potential for glacier calving front delineation from optical and radar satellite imagery. The calving front is represented as a single thin line between the ocean and the glacier, which makes the task vulnerable to inaccurate predictions. The limited availability of annotated glacier imagery leads to a lack of data diversity (not all possible combinations of different weather conditions, terminus shapes, sensors, etc. are present in the data), which exacerbates the difficulty of accurate segmentation. In this paper, we propose Attention-Multi-hooking-Deep-supervision HookNet (AMD-HookNet), a novel glacier calving front segmentation framework for synthetic aperture radar (SAR) images. The proposed method aims to enhance the feature representation capability through multiple information interactions between low-resolution and high-resolution inputs based on a two-branch U-Net. The attention mechanism, integrated into the two branch U-Net, aims to interact between the corresponding coarse and fine-grained feature maps. This allows the network to automatically adjust feature relationships, resulting in accurate pixel-classification predictions. Extensive experiments and comparisons on the challenging glacier segmentation benchmark dataset CaFFe show that our AMD-HookNet achieves a mean distance error of 438 m to the ground truth outperforming the current state of the art by 42%, which validates its effectiveness.
☆ PatchDCT: Patch Refinement for High Quality Instance Segmentation
High-quality instance segmentation has shown emerging importance in computer vision. Without any refinement, DCT-Mask directly generates high-resolution masks by compressed vectors. To further refine masks obtained by compressed vectors, we propose for the first time a compressed vector based multi-stage refinement framework. However, the vanilla combination does not bring significant gains, because changes in some elements of the DCT vector will affect the prediction of the entire mask. Thus, we propose a simple and novel method named PatchDCT, which separates the mask decoded from a DCT vector into several patches and refines each patch by the designed classifier and regressor. Specifically, the classifier is used to distinguish mixed patches from all patches, and to correct previously mispredicted foreground and background patches. In contrast, the regressor is used for DCT vector prediction of mixed patches, further refining the segmentation quality at boundary locations. Experiments on COCO show that our method achieves 2.0%, 3.2%, 4.5% AP and 3.4%, 5.3%, 7.0% Boundary AP improvements over Mask-RCNN on COCO, LVIS, and Cityscapes, respectively. It also surpasses DCT-Mask by 0.7%, 1.1%, 1.3% AP and 0.9%, 1.7%, 4.2% Boundary AP on COCO, LVIS and Cityscapes. Besides, the performance of PatchDCT is also competitive with other state-of-the-art methods.
☆ HyperSLICE: HyperBand optimised Spiral for Low-latency Interactive Cardiac Examination
BACKGROUND: Interactive cardiac magnetic resonance imaging is used for fast scan planning and MR guided interventions. However, the requirement for real-time acquisition and near real-time visualization constrains the achievable spatio-temporal resolution. PURPOSE: To improve interactive imaging resolution through optimization of undersampled spiral sampling and leveraging of deep learning for low-latency reconstruction (deep artifact suppression). POPULATION: Deep artefact suppression training data consisted of 692 breath-held CINEs. The developed interactive sequence was tested prospectively in 12 patients (10 for image evaluation, 2 during catheterization). ASSESSMENT: In simulated data, NRMSE, pSNR and SSIM of radial, uniform spiral and optimized spiral sampling were compared. In the prospective study, the optimized spiral interactive sequence was compared to conventional Cartesian real-time, and breath-hold cine imaging in terms quantitative and qualitative image metrics. RESULTS: The NRMSE, pSNR and SSIM were all statistically significantly higher in simulations of optimized spiral compared to radial and uniform spiral sampling, particularly after scan plan changes (SSIM: 0.71 vs 0.45 and 0.43). Prospectively, HyperSLICE proposed a higher spatial and temporal resolution than conventional Cartesian real-time imaging. The pipeline was demonstrated in patients during catheter pull back showing sufficiently fast reconstruction for interactive imaging. DATA CONCLUSION: HyperSLICE enables higher spatial and temporal interactive imaging. Optimizing the spiral sampling enabled better overall image quality and better handling of image transitions compared to radial and uniform spiral trajectories.
☆ 1st Place Solution for PSG competition with ECCV'22 SenseHuman Workshop
Panoptic Scene Graph (PSG) generation aims to generate scene graph representations based on panoptic segmentation instead of rigid bounding boxes. Existing PSG methods utilize one-stage paradigm which simultaneously generates scene graphs and predicts semantic segmentation masks or two-stage paradigm that first adopt an off-the-shelf panoptic segmentor, then pairwise relationship prediction between these predicted objects. One-stage approach despite having a simplified training paradigm, its segmentation results are usually under-satisfactory, while two-stage approach lacks global context and leads to low performance on relation prediction. To bridge this gap, in this paper, we propose GRNet, a Global Relation Network in two-stage paradigm, where the pre-extracted local object features and their corresponding masks are fed into a transformer with class embeddings. To handle relation ambiguity and predicate classification bias caused by long-tailed distribution, we formulate relation prediction in the second stage as a multi-class classification task with soft label. We conduct comprehensive experiments on OpenPSG dataset and achieve the state-of-art performance on the leadboard. We also show the effectiveness of our soft label strategy for long-tailed classes in ablation studies. Our code has been released in https://github.com/wangqixun/mfpsg.
comment: Tech Report
☆ Approximation of radiative transfer for surface spectral features
Remote sensing hyperspectral and more generally spectral instruments are common tools to decipher surface features in Earth and Planetary science. While linear mixture is the most common approximation for compounds detection (mineral, water, ice, etc...), the transfer of light in surface and atmospheric medium are highly non-linear. The exact simulation of non-linearities can be estimated at very high numerical cost. Here I propose a very simple non-linear form (that includes the regular linear area mixture) of radiative transfer to approximate surface spectral feature. I demonstrate that this analytical form is able to approximate the grain size and intimate mixture dependence of surface features. In addition, the same analytical form can approximate the effect of Martian mineral aerosols. Unfortunately, Earth aerosols are more complex (water droplet, water ice, soot,...) and are not expected to follow the same trend.
comment: 4 pages, 3 figures, submitted 21st october 2022 to IEEE Geoscience and Remote Sensing Letters
☆ Uncertainty Calibration and its Application to Object Detection
Image-based environment perception is an important component especially for driver assistance systems or autonomous driving. In this scope, modern neuronal networks are used to identify multiple objects as well as the according position and size information within a single frame. The performance of such an object detection model is important for the overall performance of the whole system. However, a detection model might also predict these objects under a certain degree of uncertainty. [...] In this work, we examine the semantic uncertainty (which object type?) as well as the spatial uncertainty (where is the object and how large is it?). We evaluate if the predicted uncertainties of an object detection model match with the observed error that is achieved on real-world data. In the first part of this work, we introduce the definition for confidence calibration of the semantic uncertainty in the context of object detection, instance segmentation, and semantic segmentation. We integrate additional position information in our examinations to evaluate the effect of the object's position on the semantic calibration properties. Besides measuring calibration, it is also possible to perform a post-hoc recalibration of semantic uncertainty that might have turned out to be miscalibrated. [...] The second part of this work deals with the spatial uncertainty obtained by a probabilistic detection model. [...] We review and extend common calibration methods so that it is possible to obtain parametric uncertainty distributions for the position information in a more flexible way. In the last part, we demonstrate a possible use-case for our derived calibration methods in the context of object tracking. [...] We integrate our previously proposed calibration techniques and demonstrate the usefulness of semantic and spatial uncertainty calibration in a subsequent process. [...]
comment: PhD thesis at University of Wuppertal, cite by: 'Fabian K\"uppers. "Uncertainty Calibration and its Application to Object Detection." PhD Thesis, University of Wuppertal, January 2023'
☆ COVID-19 Infection Analysis Framework using Novel Boosted CNNs and Radiological Images
COVID-19 is a new pathogen that first appeared in the human population at the end of 2019, and it can lead to novel variants of pneumonia after infection. COVID-19 is a rapidly spreading infectious disease that infects humans faster. Therefore, efficient diagnostic systems may accurately identify infected patients and thus help control their spread. In this regard, a new two-stage analysis framework is developed to analyze minute irregularities of COVID-19 infection. A novel detection Convolutional Neural Network (CNN), STM-BRNet, is developed that incorporates the Split-Transform-Merge (STM) block and channel boosting (CB) to identify COVID-19 infected CT slices in the first stage. Each STM block extracts boundary and region-smoothing-specific features for COVID-19 infection detection. Moreover, the various boosted channels are obtained by introducing the new CB and Transfer Learning (TL) concept in STM blocks to capture small illumination and texture variations of COVID-19-specific images. The COVID-19 CTs are provided with new SA-CB-BRSeg segmentation CNN for delineating infection in images in the second stage. SA-CB-BRSeg methodically utilized smoothening and heterogeneous operations in the encoder and decoder to capture simultaneously COVID-19 specific patterns that are region homogeneity, texture variation, and boundaries. Additionally, the new CB concept is introduced in the decoder of SA-CB-BRSeg by combining additional channels using TL to learn the low contrast region. The proposed STM-BRNet and SA-CB-BRSeg yield considerable achievement in accuracy: 98.01 %, Recall: 98.12%, F-score: 98.11%, and Dice Similarity: 96.396%, IOU: 98.845 % for the COVID-19 infectious region, respectively. The proposed two-stage framework significantly increased performance compared to single-phase and other reported systems and reduced the burden on the radiologists.
comment: 26 Pages, 11 Figures, 6 Tables. arXiv admin note: text overlap with arXiv:2209.10963
☆ Rethinking Out-of-distribution (OOD) Detection: Masked Image Modeling is All You Need
The core of out-of-distribution (OOD) detection is to learn the in-distribution (ID) representation, which is distinguishable from OOD samples. Previous work applied recognition-based methods to learn the ID features, which tend to learn shortcuts instead of comprehensive representations. In this work, we find surprisingly that simply using reconstruction-based methods could boost the performance of OOD detection significantly. We deeply explore the main contributors of OOD detection and find that reconstruction-based pretext tasks have the potential to provide a generally applicable and efficacious prior, which benefits the model in learning intrinsic data distributions of the ID dataset. Specifically, we take Masked Image Modeling as a pretext task for our OOD detection framework (MOOD). Without bells and whistles, MOOD outperforms previous SOTA of one-class OOD detection by 5.7%, multi-class OOD detection by 3.0%, and near-distribution OOD detection by 2.1%. It even defeats the 10-shot-per-class outlier exposure OOD detection, although we do not include any OOD samples for our detection
☆ Cluster-aware Contrastive Learning for Unsupervised Out-of-distribution Detection
Unsupervised out-of-distribution (OOD) Detection aims to separate the samples falling outside the distribution of training data without label information. Among numerous branches, contrastive learning has shown its excellent capability of learning discriminative representation in OOD detection. However, for its limited vision, merely focusing on instance-level relationship between augmented samples, it lacks attention to the relationship between samples with same semantics. Based on the classic contrastive learning, we propose Cluster-aware Contrastive Learning (CCL) framework for unsupervised OOD detection, which considers both instance-level and semantic-level information. Specifically, we study a cooperation strategy of clustering and contrastive learning to effectively extract the latent semantics and design a cluster-aware contrastive loss function to enhance OOD discriminative ability. The loss function can simultaneously pay attention to the global and local relationships by treating both the cluster centers and the samples belonging to the same cluster as positive samples. We conducted sufficient experiments to verify the effectiveness of our framework and the model achieves significant improvement on various image benchmarks.
☆ A Correction-Based Dynamic Enhancement Framework towards Underwater Detection
To assist underwater object detection for better performance, image enhancement technology is often used as a pre-processing step. However, most of the existing enhancement methods tend to pursue the visual quality of an image, instead of providing effective help for detection tasks. In fact, image enhancement algorithms should be optimized with the goal of utility improvement. In this paper, to adapt to the underwater detection tasks, we proposed a lightweight dynamic enhancement algorithm using a contribution dictionary to guide low-level corrections. Dynamic solutions are designed to capture differences in detection preferences. In addition, it can also balance the inconsistency between the contribution of correction operations and their time complexity. Experimental results in real underwater object detection tasks show the superiority of our proposed method in both generalization and real-time performance.
☆ CHiLS: Zero-Shot Image Classification with Hierarchical Label Sets
Open vocabulary models (e.g. CLIP) have shown strong performance on zero-shot classification through their ability generate embeddings for each class based on their (natural language) names. Prior work has focused on improving the accuracy of these models through prompt engineering or by incorporating a small amount of labeled downstream data (via finetuning). However, there has been little focus on improving the richness of the class names themselves, which can pose issues when class labels are coarsely-defined and uninformative. We propose Classification with Hierarchical Label Sets (or CHiLS), an alternative strategy for zero-shot classification specifically designed for datasets with implicit semantic hierarchies. CHiLS proceeds in three steps: (i) for each class, produce a set of subclasses, using either existing label hierarchies or by querying GPT-3; (ii) perform the standard zero-shot CLIP procedure as though these subclasses were the labels of interest; (iii) map the predicted subclass back to its parent to produce the final prediction. Across numerous datasets with underlying hierarchical structure, CHiLS leads to improved accuracy in situations both with and without ground-truth hierarchical information. CHiLS is simple to implement within existing CLIP pipelines and requires no additional training cost. Code is available at: https://github.com/acmi-lab/CHILS.
☆ Domain Re-Modulation for Few-Shot Generative Domain Adaptation
In this study, we investigate the task of few-shot Generative Domain Adaptation (GDA), which involves transferring a pre-trained generator from one domain to a new domain using one or a few reference images. Building upon previous research that has focused on Target-domain Consistency, Large Diversity, and Cross-domain Consistency, we conclude two additional desired properties for GDA: Memory and Domain Association. To meet these properties, we proposed a novel method Domain Re-Modulation (DoRM). Specifically, DoRM freezes the source generator and employs additional mapping and affine modules (M&A module) to capture the attributes of the target domain, resulting in a linearly combinable domain shift in style space. This allows for high-fidelity multi-domain and hybrid-domain generation by integrating multiple M&A modules in a single generator. DoRM is lightweight and easy to implement. Extensive experiments demonstrated the superior performance of DoRM on both one-shot and 10-shot GDA, both quantitatively and qualitatively. Additionally, for the first time, multi-domain and hybrid-domain generation can be achieved with a minimal storage cost by using a single model. The code will be available at https://github.com/wuyi2020/DoRM.
comment: Under Review
☆ PaRot: Patch-Wise Rotation-Invariant Network via Feature Disentanglement and Pose Restoration AAAI2023
Recent interest in point cloud analysis has led rapid progress in designing deep learning methods for 3D models. However, state-of-the-art models are not robust to rotations, which remains an unknown prior to real applications and harms the model performance. In this work, we introduce a novel Patch-wise Rotation-invariant network (PaRot), which achieves rotation invariance via feature disentanglement and produces consistent predictions for samples with arbitrary rotations. Specifically, we design a siamese training module which disentangles rotation invariance and equivariance from patches defined over different scales, e.g., the local geometry and global shape, via a pair of rotations. However, our disentangled invariant feature loses the intrinsic pose information of each patch. To solve this problem, we propose a rotation-invariant geometric relation to restore the relative pose with equivariant information for patches defined over different scales. Utilising the pose information, we propose a hierarchical module which implements intra-scale and inter-scale feature aggregation for 3D shape learning. Moreover, we introduce a pose-aware feature propagation process with the rotation-invariant relative pose information embedded. Experiments show that our disentanglement module extracts high-quality rotation-robust features and the proposed lightweight model achieves competitive results in rotated 3D object classification and part segmentation tasks. Our project page is released at: https://patchrot.github.io/.
comment: Accepted by AAAI2023
☆ Novel Fundus Image Preprocessing for Retcam Images to Improve Deep Learning Classification of Retinopathy of Prematurity
Retinopathy of Prematurity (ROP) is a potentially blinding eye disorder because of damage to the eye's retina which can affect babies born prematurely. Screening of ROP is essential for early detection and treatment. This is a laborious and manual process which requires trained physician performing dilated ophthalmological examination which can be subjective resulting in lower diagnosis success for clinically significant disease. Automated diagnostic methods can assist ophthalmologists increase diagnosis accuracy using deep learning. Several research groups have highlighted various approaches. This paper proposes the use of new novel fundus preprocessing methods using pretrained transfer learning frameworks to create hybrid models to give higher diagnosis accuracy. The evaluations show that these novel methods in comparison to traditional imaging processing contribute to higher accuracy in classifying Plus disease, Stages of ROP and Zones. We achieve accuracy of 97.65% for Plus disease, 89.44% for Stage, 90.24% for Zones
comment: 10 pages, 4 figures, 7 tables
☆ Exploiting Partial Common Information Microstructure for Multi-Modal Brain Tumor Segmentation
Learning with multiple modalities is crucial for automated brain tumor segmentation from magnetic resonance imaging data. Explicitly optimizing the common information shared among all modalities (e.g., by maximizing the total correlation) has been shown to achieve better feature representations and thus enhance the segmentation performance. However, existing approaches are oblivious to partial common information shared by subsets of the modalities. In this paper, we show that identifying such partial common information can significantly boost the discriminative power of image segmentation models. In particular, we introduce a novel concept of partial common information mask (PCI-mask) to provide a fine-grained characterization of what partial common information is shared by which subsets of the modalities. By solving a masked correlation maximization and simultaneously learning an optimal PCI-mask, we identify the latent microstructure of partial common information and leverage it in a self-attention module to selectively weight different feature representations in multi-modal data. We implement our proposed framework on the standard U-Net. Our experimental results on the Multi-modal Brain Tumor Segmentation Challenge (BraTS) datasets consistently outperform those of state-of-the-art segmentation baselines, with validation Dice similarity coefficients of 0.920, 0.897, 0.837 for the whole tumor, tumor core, and enhancing tumor on BraTS-2020.
☆ RDFNet: Regional Dynamic FISTA-Net for Spectral Snapshot Compressive Imaging
Deep convolutional neural networks have recently shown promising results in compressive spectral reconstruction. Previous methods, however, usually adopt a single mapping function for sparse representation. Considering that different regions have distinct characteristics, it is desirable to apply various mapping functions to adjust different regions' transformations dynamically. With this in mind, we first introduce a regional dynamic way of using Fast Iterative Shrinkage-Thresholding Algorithm (FISTA) to exploit regional characteristics and derive dynamic sparse representations. Then, we propose to unfold the process into a hierarchical dynamic deep network, dubbed RDFNet. The network comprises multiple regional dynamic blocks and corresponding pixel-wise adaptive soft-thresholding modules, respectively in charge of region-based dynamic mapping and pixel-wise soft-thresholding selection. The regional dynamic block guides the network to adjust the transformation domain for different regions. Equipped with the adaptive soft-thresholding, our proposed regional dynamic architecture can also learn appropriate shrinkage scale in a pixel-wise manner. Extensive experiments on both simulated and real data demonstrate that our method outperforms prior state-of-the-arts.
comment: IEEE Transactions on Computational Imaging
☆ Deep Learning for Time Series Classification and Extrinsic Regression: A Current Survey
Time Series Classification and Extrinsic Regression are important and challenging machine learning tasks. Deep learning has revolutionized natural language processing and computer vision and holds great promise in other fields such as time series analysis where the relevant features must often be abstracted from the raw data but are not known a priori. This paper surveys the current state of the art in the fast-moving field of deep learning for time series classification and extrinsic regression. We review different network architectures and training methods used for these tasks and discuss the challenges and opportunities when applying deep learning to time series data. We also summarize two critical applications of time series classification and extrinsic regression, human activity recognition and satellite earth observation.
☆ Novel Building Detection and Location Intelligence Collection in Aerial Satellite Imagery
Building structures detection and information about these buildings in aerial images is an important solution for city planning and management, land use analysis. It can be the center piece to answer important questions such as planning evacuation routes in case of an earthquake, flood management, etc. These applications rely on being able to accurately retrieve up-to-date information. Being able to accurately detect buildings in a bounding box centered on a specific latitude-longitude value can help greatly. The key challenge is to be able to detect buildings which can be commercial, industrial, hut settlements, or skyscrapers. Once we are able to detect such buildings, our goal will be to cluster and categorize similar types of buildings together.
comment: 9 pages(5 main pages, 4 auxiliary pages)
☆ Spatial Functa: Scaling Functa to ImageNet Classification and Generation
Neural fields, also known as implicit neural representations, have emerged as a powerful means to represent complex signals of various modalities. Based on this Dupont et al. (2022) introduce a framework that views neural fields as data, termed *functa*, and proposes to do deep learning directly on this dataset of neural fields. In this work, we show that the proposed framework faces limitations when scaling up to even moderately complex datasets such as CIFAR-10. We then propose *spatial functa*, which overcome these limitations by using spatially arranged latent representations of neural fields, thereby allowing us to scale up the approach to ImageNet-1k at 256x256 resolution. We demonstrate competitive performance to Vision Transformers (Steiner et al., 2022) on classification and Latent Diffusion (Rombach et al., 2022) on image generation respectively.
☆ Cooperverse: A Mobile-Edge-Cloud Framework for Universal Cooperative Perception with Mixed Connectivity and Automation
Cooperative perception (CP) is attracting increasing attention and is regarded as the core foundation to support cooperative driving automation, a potential key solution to addressing the safety, mobility, and sustainability issues of contemporary transportation systems. However, current research on CP is still at the beginning stages where a systematic problem formulation of CP is still missing, acting as the essential guideline of the system design of a CP system under real-world situations. In this paper, we formulate a universal CP system into an optimization problem and a mobile-edge-cloud framework called Cooperverse. This system addresses CP in a mixed connectivity and automation environment. A Dynamic Feature Sharing (DFS) methodology is introduced to support this CP system under certain constraints and a Random Priority Filtering (RPF) method is proposed to conduct DFS with high performance. Experiments have been conducted based on a high-fidelity CP platform, and the results show that the Cooperverse framework is effective for dynamic node engagement and the proposed DFS methodology can improve system CP performance by 14.5% and the RPF method can reduce the communication cost for mobile nodes by 90% with only 1.7% drop for average precision.
comment: 6 pages, 7 figures
☆ Studying Therapy Effects and Disease Outcomes in Silico using Artificial Counterfactual Tissue Samples
Understanding the interactions of different cell types inside the immune tumor microenvironment (iTME) is crucial for the development of immunotherapy treatments as well as for predicting their outcomes. Highly multiplexed tissue imaging (HMTI) technologies offer a tool which can capture cell properties of tissue samples by measuring expression of various proteins and storing them in separate image channels. HMTI technologies can be used to gain insights into the iTME and in particular how the iTME differs for different patient outcome groups of interest (e.g., treatment responders vs. non-responders). Understanding the systematic differences in the iTME of different patient outcome groups is crucial for developing better treatments and personalising existing treatments. However, such analyses are inherently limited by the fact that any two tissue samples vary due to a large number of factors unrelated to the outcome. Here, we present CF-HistoGAN, a machine learning framework that employs generative adversarial networks (GANs) to create artificial counterfactual tissue samples that resemble the original tissue samples as closely as possible but capture the characteristics of a different patient outcome group. Specifically, we learn to "translate" HMTI samples from one patient group to create artificial paired samples. We show that this approach allows to directly study the effects of different patient outcomes on the iTMEs of individual tissue samples. We demonstrate that CF-HistoGAN can be employed as an explorative tool for understanding iTME effects on the pixel level. Moreover, we show that our method can be used to identify statistically significant differences in the expression of different proteins between patient groups with greater sensitivity compared to conventional approaches.
☆ From CAD models to soft point cloud labels: An automatic annotation pipeline for cheaply supervised 3D semantic segmentation
We propose a fully automatic annotation scheme which takes a raw 3D point cloud with a set of fitted CAD models as input, and outputs convincing point-wise labels which can be used as cheap training data for point cloud segmentation. Compared to manual annotations, we show that our automatic labels are accurate while drastically reducing the annotation time, and eliminating the need for manual intervention or dataset-specific parameters. Our labeling pipeline outputs semantic classes and soft point-wise object scores which can either be binarized into standard one-hot-encoded labels, thresholded into weak labels with ambiguous points left unlabeled, or used directly as soft labels during training. We evaluate the label quality and segmentation performance of PointNet++ on a dataset of real industrial point clouds and Scan2CAD, a public dataset of indoor scenes. Our results indicate that reducing supervision in areas which are more difficult to label automatically is beneficial, compared to the conventional approach of naively assigning a hard "best guess" label to every point.
comment: This work has been submitted to the IEEE for possible publication
☆ Pic2Word: Mapping Pictures to Words for Zero-shot Composed Image Retrieval
In Composed Image Retrieval (CIR), a user combines a query image with text to describe their intended target. Existing methods rely on supervised learning of CIR models using labeled triplets consisting of the query image, text specification, and the target image. Labeling such triplets is expensive and hinders broad applicability of CIR. In this work, we propose to study an important task, Zero-Shot Composed Image Retrieval (ZS-CIR), whose goal is to build a CIR model without requiring labeled triplets for training. To this end, we propose a novel method, called Pic2Word, that requires only weakly labeled image-caption pairs and unlabeled image datasets to train. Unlike existing supervised CIR models, our model trained on weakly labeled or unlabeled datasets shows strong generalization across diverse ZS-CIR tasks, e.g., attribute editing, object composition, and domain conversion. Our approach outperforms several supervised CIR methods on the common CIR benchmark, CIRR and Fashion-IQ. Code will be made publicly available at https://github.com/google-research/composed_image_retrieval.
☆ Investigating Pulse-Echo Sound Speed Estimation in Breast Ultrasound with Deep Learning
Ultrasound is an adjunct tool to mammography that can quickly and safely aid physicians with diagnosing breast abnormalities. Clinical ultrasound often assumes a constant sound speed to form B-mode images for diagnosis. However, the various types of breast tissue, such as glandular, fat, and lesions, differ in sound speed. These differences can degrade the image reconstruction process. Alternatively, sound speed can be a powerful tool for identifying disease. To this end, we propose a deep-learning approach for sound speed estimation from in-phase and quadrature ultrasound signals. First, we develop a large-scale simulated ultrasound dataset that generates quasi-realistic breast tissue by modeling breast gland, skin, and lesions with varying echogenicity and sound speed. We developed a fully convolutional neural network architecture trained on a simulated dataset to produce an estimated sound speed map from inputting three complex-value in-phase and quadrature ultrasound images formed from plane-wave transmissions at separate angles. Furthermore, thermal noise augmentation is used during model optimization to enhance generalizability to real ultrasound data. We evaluate the model on simulated, phantom, and in-vivo breast ultrasound data, demonstrating its ability to accurately estimate sound speeds consistent with previously reported values in the literature. Our simulated dataset and model will be publicly available to provide a step towards accurate and generalizable sound speed estimation for pulse-echo ultrasound imaging.
♻ ☆ Deep Dependency Networks for Multi-Label Classification
We propose a simple approach which combines the strengths of probabilistic graphical models and deep learning architectures for solving the multi-label classification task, focusing specifically on image and video data. First, we show that the performance of previous approaches that combine Markov Random Fields with neural networks can be modestly improved by leveraging more powerful methods such as iterative join graph propagation, integer linear programming, and $\ell_1$ regularization-based structure learning. Then we propose a new modeling framework called deep dependency networks, which augments a dependency network, a model that is easy to train and learns more accurate dependencies but is limited to Gibbs sampling for inference, to the output layer of a neural network. We show that despite its simplicity, jointly learning this new architecture yields significant improvements in performance over the baseline neural network. In particular, our experimental evaluation on three video activity classification datasets: Charades, Textually Annotated Cooking Scenes (TACoS), and Wetlab, and three multi-label image classification datasets: MS-COCO, PASCAL VOC, and NUS-WIDE show that deep dependency networks are almost always superior to pure neural architectures that do not use dependency networks.
♻ ☆ Switchable Lightweight Anti-symmetric Processing (SLAP) with CNN Outspeeds Data Augmentation by Smaller Sample -- Application in Gomoku Reinforcement Learning
To replace data augmentation, this paper proposed a method called SLAP to intensify experience to speed up machine learning and reduce the sample size. SLAP is a model-independent protocol/function to produce the same output given different transformation variants. SLAP improved the convergence speed of convolutional neural network learning by 83% in the experiments with Gomoku game states, with only one eighth of the sample size compared with data augmentation. In reinforcement learning for Gomoku, using AlphaGo Zero/AlphaZero algorithm with data augmentation as baseline, SLAP reduced the number of training samples by a factor of 8 and achieved similar winning rate against the same evaluator, but it was not yet evident that it could speed up reinforcement learning. The benefits should at least apply to domains that are invariant to symmetry or certain transformations. As future work, SLAP may aid more explainable learning and transfer learning for domains that are not invariant to symmetry, as a small step towards artificial general intelligence.
comment: Add co-author and enrich discussion; 6 pages, 8 figures
♻ ☆ The KFIoU Loss for Rotated Object Detection ICLR 2023
Differing from the well-developed horizontal object detection area whereby the computing-friendly IoU based loss is readily adopted and well fits with the detection metrics. In contrast, rotation detectors often involve a more complicated loss based on SkewIoU which is unfriendly to gradient-based training. In this paper, we propose an effective approximate SkewIoU loss based on Gaussian modeling and Gaussian product, which mainly consists of two items. The first term is a scale-insensitive center point loss, which is used to quickly narrow the distance between the center points of the two bounding boxes. In the distance-independent second term, the product of the Gaussian distributions is adopted to inherently mimic the mechanism of SkewIoU by its definition, and show its alignment with the SkewIoU loss at trend-level within a certain distance (i.e. within 9 pixels). This is in contrast to recent Gaussian modeling based rotation detectors e.g. GWD loss and KLD loss that involve a human-specified distribution distance metric which require additional hyperparameter tuning that vary across datasets and detectors. The resulting new loss called KFIoU loss is easier to implement and works better compared with exact SkewIoU loss, thanks to its full differentiability and ability to handle the non-overlapping cases. We further extend our technique to the 3-D case which also suffers from the same issues as 2-D. Extensive results on various public datasets (2-D/3-D, aerial/text/face images) with different base detectors show the effectiveness of our approach.
comment: 18 pages, 6 figures, 8 tables, accepted by ICLR 2023, TensorFlow code: https://github.com/yangxue0827/RotationDetection, PyTorch code: https://github.com/open-mmlab/mmrotate, Jittor code: https://github.com/Jittor/JDet
♻ ☆ The Learnable Typewriter: A Generative Approach to Text Line Analysis
We present a generative document-specific approach to character analysis and recognition in text lines. Our main idea is to build on unsupervised multi-object segmentation methods and in particular those that reconstruct images based on a limited amount of visual elements, called sprites. Our approach can learn a large number of different characters and leverage line-level annotations when available. Our contribution is twofold. First, we provide the first adaptation and evaluation of a deep unsupervised multi-object segmentation approach for text line analysis. Since these methods have mainly been evaluated on synthetic data in a completely unsupervised setting, demonstrating that they can be adapted and quantitatively evaluated on real text images and that they can be trained using weak supervision are significant progresses. Second, we demonstrate the potential of our method for new applications, more specifically in the field of paleography, which studies the history and variations of handwriting, and for cipher analysis. We evaluate our approach on three very different datasets: a printed volume of the Google1000 dataset, the Copiale cipher and historical handwritten charters from the 12th and early 13th century.
comment: For the code and a quick-overview visit the project webpage at http://imagine.enpc.fr/~siglidii/learnable-typewriter
♻ ☆ High Resolution Face Editing with Masked GAN Latent Code Optimization
Face editing represents a popular research topic within the computer vision and image processing communities. While significant progress has been made recently in this area, existing solutions: (i) are still largely focused on low-resolution images, (ii) often generate editing results with visual artefacts, or (iii) lack fine-grained control and alter multiple (entangled) attributes at once, when trying to generate the desired facial semantics. In this paper, we aim to address these issues though a novel attribute editing approach called MaskFaceGAN that focuses on local attribute editing. The proposed approach is based on an optimization procedure that directly optimizes the latent code of a pre-trained (state-of-the-art) Generative Adversarial Network (i.e., StyleGAN2) with respect to several constraints that ensure: (i) preservation of relevant image content, (ii) generation of the targeted facial attributes, and (iii) spatially--selective treatment of local image areas. The constraints are enforced with the help of an (differentiable) attribute classifier and face parser that provide the necessary reference information for the optimization procedure. MaskFaceGAN is evaluated in extensive experiments on the CelebA-HQ, Helen and SiblingsDB-HQf datasets and in comparison with several state-of-the-art techniques from the literature, i.e., StarGAN, AttGAN, STGAN, and two versions of InterFaceGAN. Our experimental results show that the proposed approach is able to edit face images with respect to several local facial attributes with unprecedented image quality and at high-resolutions (1024x1024), while exhibiting considerably less problems with attribute entanglement than competing solutions. The source code is made freely available from: https://github.com/MartinPernus/MaskFaceGAN.
comment: Final ArXiv version. The paper has been accepted for publication in IEEE Transactions on Image Processing journal and will be published in 2023
♻ ☆ Object Segmentation of Cluttered Airborne LiDAR Point Clouds
Airborne topographic LiDAR is an active remote sensing technology that emits near-infrared light to map objects on the Earth's surface. Derived products of LiDAR are suitable to service a wide range of applications because of their rich three-dimensional spatial information and their capacity to obtain multiple returns. However, processing point cloud data still requires a significant effort in manual editing. Certain human-made objects are difficult to detect because of their variety of shapes, irregularly-distributed point clouds, and low number of class samples. In this work, we propose an efficient end-to-end deep learning framework to automatize the detection and segmentation of objects defined by an arbitrary number of LiDAR points surrounded by clutter. Our method is based on a light version of PointNet that achieves good performance on both object recognition and segmentation tasks. The results are tested against manually delineated power transmission towers and show promising accuracy.
comment: proceedings of the 24th International Conference of the Catalan Association for Artificial Intelligence (CCIA 2022)
♻ ☆ PandA: Unsupervised Learning of Parts and Appearances in the Feature Maps of GANs ICLR 2023
Recent advances in the understanding of Generative Adversarial Networks (GANs) have led to remarkable progress in visual editing and synthesis tasks, capitalizing on the rich semantics that are embedded in the latent spaces of pre-trained GANs. However, existing methods are often tailored to specific GAN architectures and are limited to either discovering global semantic directions that do not facilitate localized control, or require some form of supervision through manually provided regions or segmentation masks. In this light, we present an architecture-agnostic approach that jointly discovers factors representing spatial parts and their appearances in an entirely unsupervised fashion. These factors are obtained by applying a semi-nonnegative tensor factorization on the feature maps, which in turn enables context-aware local image editing with pixel-level control. In addition, we show that the discovered appearance factors correspond to saliency maps that localize concepts of interest, without using any labels. Experiments on a wide range of GAN architectures and datasets show that, in comparison to the state of the art, our method is far more efficient in terms of training time and, most importantly, provides much more accurate localized control. Our code is available at: https://github.com/james-oldfield/PandA.
comment: Accepted at ICLR 2023. Code available at: https://github.com/james-oldfield/PandA
♻ ☆ Transformation-Based Models of Video Sequences
In this work we propose a simple unsupervised approach for next frame prediction in video. Instead of directly predicting the pixels in a frame given past frames, we predict the transformations needed for generating the next frame in a sequence, given the transformations of the past frames. This leads to sharper results, while using a smaller prediction model. In order to enable a fair comparison between different video frame prediction models, we also propose a new evaluation protocol. We use generated frames as input to a classifier trained with ground truth sequences. This criterion guarantees that models scoring high are those producing sequences which preserve discriminative features, as opposed to merely penalizing any deviation, plausible or not, from the ground truth. Our proposed approach compares favourably against more sophisticated ones on the UCF-101 data set, while also being more efficient in terms of the number of parameters and computational cost.
♻ ☆ Efficient Adaptive Activation Rounding for Post-Training Quantization
Post-training quantization (PTQ) attracts increasing attention due to its convenience in deploying quantized neural networks. Rounding is the primary source of quantization error, for which previous works adopt the rounding-to-nearest scheme with a constant border of 0.5. This work demonstrates that optimizing rounding schemes can improve model accuracy. By replacing the constant border with a simple border function, we can obtain the minimal error for multiplying two numbers and eliminate the bias of its expected value, which further benefits model accuracy. Based on this insight, we approximate the border function to make the incurred overhead negligible. We also jointly optimize propagated errors and global errors. We finally propose our AQuant framework, which can learn the border function automatically. Extensive experiments show that AQuant achieves noticeable improvements compared with state-of-the-art works and pushes the accuracy of ResNet-18 up to 60.31% under the 2-bit weight and activation post-training quantization.
♻ ☆ Scanpath Prediction on Information Visualisations
We propose Unified Model of Saliency and Scanpaths (UMSS) -- a model that learns to predict visual saliency and scanpaths (i.e. sequences of eye fixations) on information visualisations. Although scanpaths provide rich information about the importance of different visualisation elements during the visual exploration process, prior work has been limited to predicting aggregated attention statistics, such as visual saliency. We present in-depth analyses of gaze behaviour for different information visualisation elements (e.g. Title, Label, Data) on the popular MASSVIS dataset. We show that while, overall, gaze patterns are surprisingly consistent across visualisations and viewers, there are also structural differences in gaze dynamics for different elements. Informed by our analyses, UMSS first predicts multi-duration element-level saliency maps, then probabilistically samples scanpaths from them. Extensive experiments on MASSVIS show that our method consistently outperforms state-of-the-art methods with respect to several, widely used scanpath and saliency evaluation metrics. Our method achieves a relative improvement in sequence score of 11.5% for scanpath prediction, and a relative improvement in Pearson correlation coefficient of up to 23.6% for saliency prediction. These results are auspicious and point towards richer user models and simulations of visual attention on visualisations without the need for any eye tracking equipment.
comment: 11 pages, 6 figures
♻ ☆ Efficient Human-in-the-loop System for Guiding DNNs Attention
Attention guidance is an approach to addressing dataset bias in deep learning, where the model relies on incorrect features to make decisions. Focusing on image classification tasks, we propose an efficient human-in-the-loop system to interactively direct the attention of classifiers to the regions specified by users, thereby reducing the influence of co-occurrence bias and improving the transferability and interpretability of a DNN. Previous approaches for attention guidance require the preparation of pixel-level annotations and are not designed as interactive systems. We present a new interactive method to allow users to annotate images with simple clicks, and study a novel active learning strategy to significantly reduce the number of annotations. We conducted both a numerical evaluation and a user study to evaluate the proposed system on multiple datasets. Compared to the existing non-active-learning approach which usually relies on huge amounts of polygon-based segmentation masks to fine-tune or train the DNNs, our system can save lots of labor and money and obtain a fine-tuned network that works better even when the dataset is biased. The experiment results indicate that the proposed system is efficient, reasonable, and reliable.
comment: 13 pages, 11 figures, proceeding of ACM IUI 2023, video https://youtu.be/2MD-z6vXKJ4
♻ ☆ H2RBox: Horizontal Box Annotation is All You Need for Oriented Object Detection ICLR 2023
Oriented object detection emerges in many applications from aerial images to autonomous driving, while many existing detection benchmarks are annotated with horizontal bounding box only which is also less costive than fine-grained rotated box, leading to a gap between the readily available training corpus and the rising demand for oriented object detection. This paper proposes a simple yet effective oriented object detection approach called H2RBox merely using horizontal box annotation for weakly-supervised training, which closes the above gap and shows competitive performance even against those trained with rotated boxes. The cores of our method are weakly- and self-supervised learning, which predicts the angle of the object by learning the consistency of two different views. To our best knowledge, H2RBox is the first horizontal box annotation-based oriented object detector. Compared to an alternative i.e. horizontal box-supervised instance segmentation with our post adaption to oriented object detection, our approach is not susceptible to the prediction quality of mask and can perform more robustly in complex scenes containing a large number of dense objects and outliers. Experimental results show that H2RBox has significant performance and speed advantages over horizontal box-supervised instance segmentation methods, as well as lower memory requirements. While compared to rotated box-supervised oriented object detectors, our method shows very close performance and speed. The source code is available at PyTorch-based \href{https://github.com/yangxue0827/h2rbox-mmrotate}{MMRotate} and Jittor-based \href{https://github.com/yangxue0827/h2rbox-jittor}{JDet}.
comment: 15 pages, 6 figures, 7 tables, accepted by ICLR 2023, the source code is available at https://github.com/yangxue0827/h2rbox-mmrotate and https://github.com/yangxue0827/h2rbox-jittor
♻ ☆ Fruit Ripeness Classification: a Survey
Fruit is a key crop in worldwide agriculture feeding millions of people. The standard supply chain of fruit products involves quality checks to guarantee freshness, taste, and, most of all, safety. An important factor that determines fruit quality is its stage of ripening. This is usually manually classified by experts in the field, which makes it a labor-intensive and error-prone process. Thus, there is an arising need for automation in the process of fruit ripeness classification. Many automatic methods have been proposed that employ a variety of feature descriptors for the food item to be graded. Machine learning and deep learning techniques dominate the top-performing methods. Furthermore, deep learning can operate on raw data and thus relieve the users from having to compute complex engineered features, which are often crop-specific. In this survey, we review the latest methods proposed in the literature to automatize fruit ripeness classification, highlighting the most common feature descriptors they operate on.
♻ ☆ A Theoretical Framework for AI Models Explainability
EXplainable Artificial Intelligence (XAI) is a vibrant research topic in the artificial intelligence community, with growing interest across methods and domains. Much has been written about the subject, yet XAI still lacks shared terminology and a framework capable of providing structural soundness to explanations. In our work, we address these issues by proposing a novel definition of explanation that is a synthesis of what can be found in the literature. We recognize that explanations are not atomic but the combination of evidence stemming from the model and its input-output mapping, and the human interpretation of this evidence. Furthermore, we fit explanations into the properties of faithfulness (i.e., the explanation being a true description of the model's inner workings and decision-making process) and plausibility (i.e., how much the explanation looks convincing to the user). Using our proposed theoretical framework simplifies how these properties are operationalized and it provides new insight into common explanation methods that we analyze as case studies.
♻ ☆ Human alignment of neural network representations ICLR 2023
Today's computer vision models achieve human or near-human level performance across a wide variety of vision tasks. However, their architectures, data, and learning algorithms differ in numerous ways from those that give rise to human vision. In this paper, we investigate the factors that affect the alignment between the representations learned by neural networks and human mental representations inferred from behavioral responses. We find that model scale and architecture have essentially no effect on the alignment with human behavioral responses, whereas the training dataset and objective function both have a much larger impact. These findings are consistent across three datasets of human similarity judgments collected using two different tasks. Linear transformations of neural network representations learned from behavioral responses from one dataset substantially improve alignment with human similarity judgments on the other two datasets. In addition, we find that some human concepts such as food and animals are well-represented by neural networks whereas others such as royal or sports-related objects are not. Overall, although models trained on larger, more diverse datasets achieve better alignment with humans than models trained on ImageNet alone, our results indicate that scaling alone is unlikely to be sufficient to train neural networks with conceptual representations that match those used by humans.
comment: Accepted for publication at ICLR 2023
♻ ☆ Contracting Skeletal Kinematic Embeddings for Anomaly Detection
Detecting the anomaly of human behavior is paramount to timely recognizing endangering situations, such as street fights or elderly falls. However, anomaly detection is complex, since anomalous events are rare and because it is an open set recognition task, i.e., what is anomalous at inference has not been observed at training. We propose COSKAD, a novel model which encodes skeletal human motion by an efficient graph convolutional network and learns to COntract SKeletal kinematic embeddings onto a latent hypersphere of minimum volume for Anomaly Detection. We propose and analyze three latent space designs for COSKAD: the commonly-adopted Euclidean, and the new spherical-radial and hyperbolic volumes. All three variants outperform the state-of-the-art, including video-based techniques, on the ShangaiTechCampus, the Avenue, and on the most recent UBnormal dataset, for which we contribute novel skeleton annotations and the selection of human-related videos. The source code and dataset will be released upon acceptance.
comment: Submitted to Patter Recognition Journal
♻ ☆ Re-parameterizing Your Optimizers rather than Architectures ICLR 2023
The well-designed structures in neural networks reflect the prior knowledge incorporated into the models. However, though different models have various priors, we are used to training them with model-agnostic optimizers such as SGD. In this paper, we propose to incorporate model-specific prior knowledge into optimizers by modifying the gradients according to a set of model-specific hyper-parameters. Such a methodology is referred to as Gradient Re-parameterization, and the optimizers are named RepOptimizers. For the extreme simplicity of model structure, we focus on a VGG-style plain model and showcase that such a simple model trained with a RepOptimizer, which is referred to as RepOpt-VGG, performs on par with or better than the recent well-designed models. From a practical perspective, RepOpt-VGG is a favorable base model because of its simple structure, high inference speed and training efficiency. Compared to Structural Re-parameterization, which adds priors into models via constructing extra training-time structures, RepOptimizers require no extra forward/backward computations and solve the problem of quantization. We hope to spark further research beyond the realms of model structure design. Code and models \url{https://github.com/DingXiaoH/RepOptimizers}.
comment: ICLR 2023
♻ ☆ Show me your NFT and I tell you how it will perform: Multimodal representation learning for NFT selling price prediction
Non-Fungible Tokens (NFTs) represent deeds of ownership, based on blockchain technologies and smart contracts, of unique crypto assets on digital art forms (e.g., artworks or collectibles). In the spotlight after skyrocketing in 2021, NFTs have attracted the attention of crypto enthusiasts and investors intent on placing promising investments in this profitable market. However, the NFT financial performance prediction has not been widely explored to date. In this work, we address the above problem based on the hypothesis that NFT images and their textual descriptions are essential proxies to predict the NFT selling prices. To this purpose, we propose MERLIN, a novel multimodal deep learning framework designed to train Transformer-based language and visual models, along with graph neural network models, on collections of NFTs' images and texts. A key aspect in MERLIN is its independence on financial features, as it exploits only the primary data a user interested in NFT trading would like to deal with, i.e., NFT images and textual descriptions. By learning dense representations of such data, a price-category classification task is performed by MERLIN models, which can also be tuned according to user preferences in the inference phase to mimic different risk-return investment profiles. Experimental evaluation on a publicly available dataset has shown that MERLIN models achieve significant performances according to several financial assessment criteria, fostering profitable investments, and also beating baseline machine-learning classifiers based on financial features.
comment: Accepted paper at The ACM Web Conference 2023, April 30--May 04, 2023, Austin, Texas, USA
♻ ☆ LAP: An Attention-Based Module for Faithful Interpretation and Knowledge Injection in Convolutional Neural Networks
Despite the state-of-the-art performance of deep convolutional neural networks, they are susceptible to bias and malfunction in unseen situations. The complex computation behind their reasoning is not sufficiently human-understandable to develop trust. External explainer methods have tried to interpret the network decisions in a human-understandable way, but they are accused of fallacies due to their assumptions and simplifications. On the other side, the inherent self-interpretability of models, while being more robust to the mentioned fallacies, cannot be applied to the already trained models. In this work, we propose a new attention-based pooling layer, called Local Attention Pooling (LAP), that accomplishes self-interpretability and the possibility for knowledge injection while improving the model's performance. Moreover, several weakly-supervised knowledge injection methodologies are provided to enhance the process of training. We verified our claims by evaluating several LAP-extended models on three different datasets, including Imagenet. The proposed framework offers more valid human-understandable and more faithful-to-the-model interpretations than the commonly used white-box explainer methods.
♻ ☆ Dual PatchNorm
We propose Dual PatchNorm: two Layer Normalization layers (LayerNorms), before and after the patch embedding layer in Vision Transformers. We demonstrate that Dual PatchNorm outperforms the result of exhaustive search for alternative LayerNorm placement strategies in the Transformer block itself. In our experiments, incorporating this trivial modification, often leads to improved accuracy over well-tuned Vision Transformers and never hurts.
♻ ☆ Get3DHuman: Lifting StyleGAN-Human into a 3D Generative Model using Pixel-aligned Reconstruction Priors
Fast generation of high-quality 3D digital humans is important to a vast number of applications ranging from entertainment to professional concerns. Recent advances in differentiable rendering have enabled the training of 3D generative models without requiring 3D ground truths. However, the quality of the generated 3D humans still has much room to improve in terms of both fidelity and diversity. In this paper, we present Get3DHuman, a novel 3D human framework that can significantly boost the realism and diversity of the generated outcomes by only using a limited budget of 3D ground-truth data. Our key observation is that the 3D generator can profit from human-related priors learned through 2D human generators and 3D reconstructors. Specifically, we bridge the latent space of Get3DHuman with that of StyleGAN-Human via a specially-designed prior network, where the input latent code is mapped to the shape and texture feature volumes spanned by the pixel-aligned 3D reconstructor. The outcomes of the prior network are then leveraged as the supervisory signals for the main generator network. To ensure effective training, we further propose three tailored losses applied to the generated feature volumes and the intermediate feature maps. Extensive experiments demonstrate that Get3DHuman greatly outperforms the other state-of-the-art approaches and can support a wide range of applications including shape interpolation, shape re-texturing, and single-view reconstruction through latent inversion.
♻ ☆ Tuning-free Plug-and-Play Hyperspectral Image Deconvolution with Deep Priors
Deconvolution is a widely used strategy to mitigate the blurring and noisy degradation of hyperspectral images~(HSI) generated by the acquisition devices. This issue is usually addressed by solving an ill-posed inverse problem. While investigating proper image priors can enhance the deconvolution performance, it is not trivial to handcraft a powerful regularizer and to set the regularization parameters. To address these issues, in this paper we introduce a tuning-free Plug-and-Play (PnP) algorithm for HSI deconvolution. Specifically, we use the alternating direction method of multipliers (ADMM) to decompose the optimization problem into two iterative sub-problems. A flexible blind 3D denoising network (B3DDN) is designed to learn deep priors and to solve the denoising sub-problem with different noise levels. A measure of 3D residual whiteness is then investigated to adjust the penalty parameters when solving the quadratic sub-problems, as well as a stopping criterion. Experimental results on both simulated and real-world data with ground-truth demonstrate the superiority of the proposed method.
comment: IEEE Trans. Geosci. Remote sens., to be published. Manuscript submitted Jun. 30, 2022; revised Oct. 25, 2022, and Dec. 06, 2022; and accepted Feb. 02, 2023
♻ ☆ Multi-Domain Long-Tailed Learning by Augmenting Disentangled Representations
There is an inescapable long-tailed class-imbalance issue in many real-world classification problems. Existing long-tailed classification methods focus on the single-domain setting, where all examples are drawn from the same distribution. However, real-world scenarios often involve multiple domains with distinct imbalanced class distributions. We study this multi-domain long-tailed learning problem and aim to produce a model that generalizes well across all classes and domains. Towards that goal, we introduce TALLY, which produces invariant predictors by balanced augmenting hidden representations over domains and classes. Built upon a proposed selective balanced sampling strategy, TALLY achieves this by mixing the semantic representation of one example with the domain-associated nuisances of another, producing a new representation for use as data augmentation. To improve the disentanglement of semantic representations, TALLY further utilizes a domain-invariant class prototype that averages out domain-specific effects. We evaluate TALLY on four long-tailed variants of classical domain generalization benchmarks and two real-world imbalanced multi-domain datasets. The results indicate that TALLY consistently outperforms other state-of-the-art methods in both subpopulation shift and domain shift.
♻ ☆ Semantic-Aligned Matching for Enhanced DETR Convergence and Multi-Scale Feature Fusion
The recently proposed DEtection TRansformer (DETR) has established a fully end-to-end paradigm for object detection. However, DETR suffers from slow training convergence, which hinders its applicability to various detection tasks. We observe that DETR's slow convergence is largely attributed to the difficulty in matching object queries to relevant regions due to the unaligned semantics between object queries and encoded image features. With this observation, we design Semantic-Aligned-Matching DETR++ (SAM-DETR++) to accelerate DETR's convergence and improve detection performance. The core of SAM-DETR++ is a plug-and-play module that projects object queries and encoded image features into the same feature embedding space, where each object query can be easily matched to relevant regions with similar semantics. Besides, SAM-DETR++ searches for multiple representative keypoints and exploits their features for semantic-aligned matching with enhanced representation capacity. Furthermore, SAM-DETR++ can effectively fuse multi-scale features in a coarse-to-fine manner on the basis of the designed semantic-aligned matching. Extensive experiments show that the proposed SAM-DETR++ achieves superior convergence speed and competitive detection accuracy. Additionally, as a plug-and-play method, SAM-DETR++ can complement existing DETR convergence solutions with even better performance, achieving 44.8% AP with merely 12 training epochs and 49.1% AP with 50 training epochs on COCO val2017 with ResNet-50. Codes are available at https://github.com/ZhangGongjie/SAM-DETR .
♻ ☆ ERA-Solver: Error-Robust Adams Solver for Fast Sampling of Diffusion Probabilistic Models
Though denoising diffusion probabilistic models (DDPMs) have achieved remarkable generation results, the low sampling efficiency of DDPMs still limits further applications. Since DDPMs can be formulated as diffusion ordinary differential equations (ODEs), various fast sampling methods can be derived from solving diffusion ODEs. However, we notice that previous sampling methods with fixed analytical form are not robust with the error in the noise estimated from pretrained diffusion models. In this work, we construct an error-robust Adams solver (ERA-Solver), which utilizes the implicit Adams numerical method that consists of a predictor and a corrector. Different from the traditional predictor based on explicit Adams methods, we leverage a Lagrange interpolation function as the predictor, which is further enhanced with an error-robust strategy to adaptively select the Lagrange bases with lower error in the estimated noise. Experiments on Cifar10, LSUN-Church, and LSUN-Bedroom datasets demonstrate that our proposed ERA-Solver achieves 5.14, 9.42, and 9.69 Fenchel Inception Distance (FID) for image generation, with only 10 network evaluations.
comment: 16 pages, 12 figures
♻ ☆ Global Context Vision Transformers
We propose global context vision transformer (GC ViT), a novel architecture that enhances parameter and compute utilization for computer vision tasks. The core of the novel model are global context self-attention modules, joint with standard local self-attention, to effectively yet efficiently model both long and short-range spatial interactions, as an alternative to complex operations such as an attention masks or local windows shifting. While the local self-attention modules are responsible for modeling short-range information, the global query tokens are shared across all global self-attention modules to interact with local key and values. In addition, we address the lack of inductive bias in ViTs and improve the modeling of inter-channel dependencies by proposing a novel downsampler which leverages a parameter-efficient fused inverted residual block. The proposed GC ViT achieves new state-of-the-art performance across image classification, object detection and semantic segmentation tasks. On ImageNet-1K dataset for classification, GC ViT models with 51M, 90M and 201M parameters achieve 84.3%, 84.9% and 85.6% Top-1 accuracy, respectively, surpassing comparably-sized prior art such as CNN-based ConvNeXt and ViT-based Swin Transformer. Pre-trained GC ViT backbones in downstream tasks of object detection, instance segmentation, and semantic segmentation on MS COCO and ADE20K datasets outperform prior work consistently, sometimes by large margins.
comment: Tech report
♻ ☆ Galaxy Image Deconvolution for Weak Gravitational Lensing with Unrolled Plug-and-Play ADMM
Removing optical and atmospheric blur from galaxy images significantly improves galaxy shape measurements for weak gravitational lensing and galaxy evolution studies. This ill-posed linear inverse problem is usually solved with deconvolution algorithms enhanced by regularisation priors or deep learning. We introduce a so-called "physics-based deep learning" approach to the Point Spread Function (PSF) deconvolution problem in galaxy surveys. We apply algorithm unrolling and the Plug-and-Play technique to the Alternating Direction Method of Multipliers (ADMM), in which a neural network learns appropriate hyperparameters and denoising priors from simulated galaxy images. We characterise the time-performance trade-off of several methods for galaxies of differing brightness levels as well as our method's robustness to systematic PSF errors and ablations. We show an improvement in reduced shear ellipticity error of 38.6% (SNR=20)/45.0% (SNR=200) compared to classic methods and 7.4% (SNR=20)/33.2% (SNR=200) compared to modern methods.
♻ ☆ Single Shot Self-Reliant Scene Text Spotter by Decoupled yet Collaborative Detection and Recognition
Typical text spotters follow the two-stage spotting paradigm which detects the boundary for a text instance first and then performs text recognition within the detected regions. Despite the remarkable progress of such spotting paradigm, an important limitation is that the performance of text recognition depends heavily on the precision of text detection, resulting in the potential error propagation from detection to recognition. In this work, we propose the single shot Self-Reliant Scene Text Spotter v2 (SRSTS v2), which circumvents this limitation by decoupling recognition from detection while optimizing two tasks collaboratively. Specifically, our SRSTS v2 samples representative feature points around each potential text instance, and conducts both text detection and recognition in parallel guided by these sampled points. Thus, the text recognition is no longer dependent on detection, thereby alleviating the error propagation from detection to recognition. Moreover, the sampling module is learned under the supervision from both detection and recognition, which allows for the collaborative optimization and mutual enhancement between two tasks. Benefiting from such sampling-driven concurrent spotting framework, our approach is able to recognize the text instances correctly even if the precise text boundaries are challenging to detect. Extensive experiments on four benchmarks demonstrate that our method compares favorably to state-of-the-art spotters.
♻ ☆ PriorLane: A Prior Knowledge Enhanced Lane Detection Approach Based on Transformer ICRA 2023
Lane detection is one of the fundamental modules in self-driving. In this paper we employ a transformer-only method for lane detection, thus it could benefit from the blooming development of fully vision transformer and achieve the state-of-the-art (SOTA) performance on both CULane and TuSimple benchmarks, by fine-tuning the weight fully pre-trained on large datasets. More importantly, this paper proposes a novel and general framework called PriorLane, which is used to enhance the segmentation performance of the fully vision transformer by introducing the low-cost local prior knowledge. Specifically, PriorLane utilizes an encoder-only transformer to fuse the feature extracted by a pre-trained segmentation model with prior knowledge embeddings. Note that a Knowledge Embedding Alignment (KEA) module is adapted to enhance the fusion performance by aligning the knowledge embedding. Extensive experiments on our Zjlab dataset show that PriorLane outperforms SOTA lane detection methods by a 2.82% mIoU when prior knowledge is employed.
comment: Accepted by ICRA 2023
♻ ☆ CHIMLE: Conditional Hierarchical IMLE for Multimodal Conditional Image Synthesis
A persistent challenge in conditional image synthesis has been to generate diverse output images from the same input image despite only one output image being observed per input image. GAN-based methods are prone to mode collapse, which leads to low diversity. To get around this, we leverage Implicit Maximum Likelihood Estimation (IMLE) which can overcome mode collapse fundamentally. IMLE uses the same generator as GANs but trains it with a different, non-adversarial objective which ensures each observed image has a generated sample nearby. Unfortunately, to generate high-fidelity images, prior IMLE-based methods require a large number of samples, which is expensive. In this paper, we propose a new method to get around this limitation, which we dub Conditional Hierarchical IMLE (CHIMLE), which can generate high-fidelity images without requiring many samples. We show CHIMLE significantly outperforms the prior best IMLE, GAN and diffusion-based methods in terms of image fidelity and mode coverage across four tasks, namely night-to-day, 16x single image super-resolution, image colourization and image decompression. Quantitatively, our method improves Fr\'echet Inception Distance (FID) by 36.9% on average compared to the prior best IMLE-based method, and by 27.5% on average compared to the best non-IMLE-based general-purpose methods.
comment: More results and code are available on the project website at https://niopeng.github.io/CHIMLE/
♻ ☆ Convolutional Neural Generative Coding: Scaling Predictive Coding to Natural Images
In this work, we develop convolutional neural generative coding (Conv-NGC), a generalization of predictive coding to the case of convolution/deconvolution-based computation. Specifically, we concretely implement a flexible neurobiologically-motivated algorithm that progressively refines latent state feature maps in order to dynamically form a more accurate internal representation/reconstruction model of natural images. The performance of the resulting sensory processing system is evaluated on complex datasets such as Color-MNIST, CIFAR-10, and Street House View Numbers (SVHN). We study the effectiveness of our brain-inspired model on the tasks of reconstruction and image denoising and find that it is competitive with convolutional auto-encoding systems trained by backpropagation of errors and outperforms them with respect to out-of-distribution reconstruction (including the full 90k CINIC-10 test set).
comment: Revisions/updates, expanded appendix
♻ ☆ An Uncertainty-aware Loss Function for Training Neural Networks with Calibrated Predictions
Uncertainty quantification of machine learning and deep learning methods plays an important role in enhancing trust to the obtained result. In recent years, a numerous number of uncertainty quantification methods have been introduced. Monte Carlo dropout (MC-Dropout) is one of the most well-known techniques to quantify uncertainty in deep learning methods. In this study, we propose two new loss functions by combining cross entropy with Expected Calibration Error (ECE) and Predictive Entropy (PE). The obtained results clearly show that the new proposed loss functions lead to having a calibrated MC-Dropout method. Our results confirmed the great impact of the new hybrid loss functions for minimising the overlap between the distributions of uncertainty estimates for correct and incorrect predictions without sacrificing the model's overall performance.
comment: 11 pages, 6 figures, 2 tables
♻ ☆ RegMixup: Mixup as a Regularizer Can Surprisingly Improve Accuracy and Out Distribution Robustness
We show that the effectiveness of the well celebrated Mixup [Zhang et al., 2018] can be further improved if instead of using it as the sole learning objective, it is utilized as an additional regularizer to the standard cross-entropy loss. This simple change not only provides much improved accuracy but also significantly improves the quality of the predictive uncertainty estimation of Mixup in most cases under various forms of covariate shifts and out-of-distribution detection experiments. In fact, we observe that Mixup yields much degraded performance on detecting out-of-distribution samples possibly, as we show empirically, because of its tendency to learn models that exhibit high-entropy throughout; making it difficult to differentiate in-distribution samples from out-distribution ones. To show the efficacy of our approach (RegMixup), we provide thorough analyses and experiments on vision datasets (ImageNet & CIFAR-10/100) and compare it with a suite of recent approaches for reliable uncertainty estimation.
comment: 22 pages, 18 figures
♻ ☆ Realistic Defocus Blur for Multiplane Computer-Generated Holography
This paper introduces a new multiplane CGH computation method to reconstruct artefact-free high-quality holograms with natural-looking defocus blur. Our method introduces a new targeting scheme and a new loss function. While the targeting scheme accounts for defocused parts of the scene at each depth plane, the new loss function analyzes focused and defocused parts separately in reconstructed images. Our method support phase-only CGH calculations using various iterative (e.g., Gerchberg-Saxton, Gradient Descent) and non-iterative (e.g., Double Phase) CGH techniques. We achieve our best image quality using a modified gradient descent-based optimization recipe where we introduce a constraint inspired by the double phase method. We validate our method experimentally using our proof-of-concept holographic display, comparing various algorithms, including multi-depth scenes with sparse and dense contents.
comment: 16 pages in total, first 9 pages are for the manuscript, remaining pages are for supplementary. For more visit: https://complightlab.com/publications/realistic_defocus_cgh For our codebase visit https://github.com/complight/realistic_defocus
♻ ☆ Euclidean and Affine Curve Reconstruction
We consider practical aspects of reconstructing planar curves with prescribed Euclidean or affine curvatures. These curvatures are invariant under the special Euclidean group and the equi-affine groups, respectively, and play an important role in computer vision and shape analysis. We discuss and implement algorithms for such reconstruction, and give estimates on how close reconstructed curves are relative to the closeness of their curvatures in appropriate metrics. Several illustrative examples are provided.
comment: This paper is a result of an REU project conducted at the North Carolina State University in the Summer and Fall 2020. This version, with improved quality of presentation and figures, is accepted to "Involve" https://msp.org/involve/about/journal/about.html
♻ ☆ Survey: Image Mixing and Deleting for Data Augmentation
Neural networks are prone to overfitting and memorizing data patterns. To avoid over-fitting and enhance their generalization and performance, various methods have been suggested in the literature, including dropout, regularization, label smoothing, etc. One such method is augmentation which introduces different types of corruption in the data to prevent the model from overfitting and to memorize patterns present in the data. A sub-area of data augmentation is image mixing and deleting. This specific type of augmentation either deletes image regions or mixes two images to hide or make particular characteristics of images confusing for the network, forcing it to emphasize the overall structure of the object in an image. Models trained with this approach have proven to perform and generalize well compared to those trained without image mixing or deleting. An added benefit that comes with this method of training is robustness against image corruption. Due to its low computational cost and recent success, researchers have proposed many image mixing and deleting techniques. We furnish an in-depth survey of image mixing and deleting techniques and provide categorization via their most distinguishing features. We initiate our discussion with some fundamental relevant concepts. Next, we present essentials, such as each category's strengths and limitations, describing their working mechanism, basic formulations, and applications. We also discuss the general challenges and recommend possible future research directions for image mixing and deleting data augmentation techniques. Datasets and codes for evaluation are publicly available here.
Information Retrieval 10
☆ LexLIP: Lexicon-Bottlenecked Language-Image Pre-Training for Large-Scale Image-Text Retrieval
Image-text retrieval (ITR) is a task to retrieve the relevant images/texts, given the query from another modality. The conventional dense retrieval paradigm relies on encoding images and texts into dense representations using dual-stream encoders, however, it faces challenges with low retrieval speed in large-scale retrieval scenarios. In this work, we propose the lexicon-weighting paradigm, where sparse representations in vocabulary space are learned for images and texts to take advantage of the bag-of-words models and efficient inverted indexes, resulting in significantly reduced retrieval latency. A crucial gap arises from the continuous nature of image data, and the requirement for a sparse vocabulary space representation. To bridge this gap, we introduce a novel pre-training framework, Lexicon-Bottlenecked Language-Image Pre-Training (LexLIP), that learns importance-aware lexicon representations. This framework features lexicon-bottlenecked modules between the dual-stream encoders and weakened text decoders, allowing for constructing continuous bag-of-words bottlenecks to learn lexicon-importance distributions. Upon pre-training with same-scale data, our LexLIP achieves state-of-the-art performance on two benchmark ITR datasets, MSCOCO and Flickr30k. Furthermore, in large-scale retrieval scenarios, LexLIP outperforms CLIP with a 5.5 ~ 221.3X faster retrieval speed and 13.2 ~ 48.8X less index storage memory.
☆ Divide and Conquer: Towards Better Embedding-based Retrieval for Recommender Systems From a Multi-task Perspective WWW'23
Embedding-based retrieval (EBR) methods are widely used in modern recommender systems thanks to its simplicity and effectiveness. However, along the journey of deploying and iterating on EBR in production, we still identify some fundamental issues in existing methods. First, when dealing with large corpus of candidate items, EBR models often have difficulties in balancing the performance on distinguishing highly relevant items (positives) from both irrelevant ones (easy negatives) and from somewhat related yet not competitive ones (hard negatives). Also, we have little control in the diversity and fairness of the retrieval results because of the ``greedy'' nature of nearest vector search. These issues compromise the performance of EBR methods in large-scale industrial scenarios. This paper introduces a simple and proven-in-production solution to overcome these issues. The proposed solution takes a divide-and-conquer approach: the whole set of candidate items are divided into multiple clusters and we run EBR to retrieve relevant candidates from each cluster in parallel; top candidates from each cluster are then combined by some controllable merging strategies. This approach allows our EBR models to only concentrate on discriminating positives from mostly hard negatives. It also enables further improvement from a multi-tasking learning (MTL) perspective: retrieval problems within each cluster can be regarded as individual tasks; inspired by recent successes in prompting and prefix-tuning, we propose an efficient task adaption technique further boosting the retrieval performance within each cluster with negligible overheads.
comment: To appear in WWW'23 (Industry Track)
☆ Hybrid Contrastive Constraints for Multi-Scenario Ad Ranking
Multi-scenario ad ranking aims at leveraging the data from multiple domains or channels for training a unified ranking model to improve the performance at each individual scenario. Although the research on this task has made important progress, it still lacks the consideration of cross-scenario relations, thus leading to limitation in learning capability and difficulty in interrelation modeling. In this paper, we propose a Hybrid Contrastive Constrained approach (HC^2) for multi-scenario ad ranking. To enhance the modeling of data interrelation, we elaborately design a hybrid contrastive learning approach to capture commonalities and differences among multiple scenarios. The core of our approach consists of two elaborated contrastive losses, namely generalized and individual contrastive loss, which aim at capturing common knowledge and scenario-specific knowledge, respectively. To adapt contrastive learning to the complex multi-scenario setting, we propose a series of important improvements. For generalized contrastive loss, we enhance contrastive learning by extending the contrastive samples (label-aware and diffusion noise enhanced contrastive samples) and reweighting the contrastive samples (reciprocal similarity weighting). For individual contrastive loss, we use the strategies of dropout-based augmentation and {cross-scenario encoding} for generating meaningful positive and negative contrastive samples, respectively. Extensive experiments on both offline evaluation and online test have demonstrated the effectiveness of the proposed HC$^2$ by comparing it with a number of competitive baselines.
comment: 10 pages, 5 figures
☆ RLTP: Reinforcement Learning to Pace for Delayed Impression Modeling in Preloaded Ads
To increase brand awareness, many advertisers conclude contracts with advertising platforms to purchase traffic and then deliver advertisements to target audiences. In a whole delivery period, advertisers usually desire a certain impression count for the ads, and they also expect that the delivery performance is as good as possible (e.g., obtaining high click-through rate). Advertising platforms employ pacing algorithms to satisfy the demands via adjusting the selection probabilities to traffic requests in real-time. However, the delivery procedure is also affected by the strategies from publishers, which cannot be controlled by advertising platforms. Preloading is a widely used strategy for many types of ads (e.g., video ads) to make sure that the response time for displaying after a traffic request is legitimate, which results in delayed impression phenomenon. Traditional pacing algorithms cannot handle the preloading nature well because they rely on immediate feedback signals, and may fail to guarantee the demands from advertisers. In this paper, we focus on a new research problem of impression pacing for preloaded ads, and propose a Reinforcement Learning To Pace framework RLTP. It learns a pacing agent that sequentially produces selection probabilities in the whole delivery period. To jointly optimize the two objectives of impression count and delivery performance, RLTP employs tailored reward estimator to satisfy the guaranteed impression count, penalize the over-delivery and maximize the traffic value. Experiments on large-scale industrial datasets verify that RLTP outperforms baseline pacing algorithms by a large margin. We have deployed the RLTP framework online to our advertising platform, and results show that it achieves significant uplift to core metrics including delivery completion rate and click-through rate.
comment: The first two authors contributed equally
☆ Recommender Systems: A Primer
Personalized recommendations have become a common feature of modern online services, including most major e-commerce sites, media platforms and social networks. Today, due to their high practical relevance, research in the area of recommender systems is flourishing more than ever. However, with the new application scenarios of recommender systems that we observe today, constantly new challenges arise as well, both in terms of algorithmic requirements and with respect to the evaluation of such systems. In this paper, we first provide an overview of the traditional formulation of the recommendation problem. We then review the classical algorithmic paradigms for item retrieval and ranking and elaborate how such systems can be evaluated. Afterwards, we discuss a number of recent developments in recommender systems research, including research on session-based recommendation, biases in recommender systems, and questions regarding the impact and value of recommender systems in practice.
☆ Learned Accelerator Framework for Angular-Distance-Based High-Dimensional DBSCAN EDBT 2023
Density-based clustering is a commonly used tool in data science. Today many data science works are utilizing high-dimensional neural embeddings. However, traditional density-based clustering techniques like DBSCAN have a degraded performance on high-dimensional data. In this paper, we propose LAF, a generic learned accelerator framework to speed up the original DBSCAN and the sampling-based variants of DBSCAN on high-dimensional data with angular distance metric. This framework consists of a learned cardinality estimator and a post-processing module. The cardinality estimator can fast predict whether a data point is core or not to skip unnecessary range queries, while the post-processing module detects the false negative predictions and merges the falsely separated clusters. The evaluation shows our LAF-enhanced DBSCAN method outperforms the state-of-the-art efficient DBSCAN variants on both efficiency and quality.
comment: Accepted by EDBT 2023
☆ PIER: Permutation-Level Interest-Based End-to-End Re-ranking Framework in E-commerce
Re-ranking draws increased attention on both academics and industries, which rearranges the ranking list by modeling the mutual influence among items to better meet users' demands. Many existing re-ranking methods directly take the initial ranking list as input, and generate the optimal permutation through a well-designed context-wise model, which brings the evaluation-before-reranking problem. Meanwhile, evaluating all candidate permutations brings unacceptable computational costs in practice. Thus, to better balance efficiency and effectiveness, online systems usually use a two-stage architecture which uses some heuristic methods such as beam-search to generate a suitable amount of candidate permutations firstly, which are then fed into the evaluation model to get the optimal permutation. However, existing methods in both stages can be improved through the following aspects. As for generation stage, heuristic methods only use point-wise prediction scores and lack an effective judgment. As for evaluation stage, most existing context-wise evaluation models only consider the item context and lack more fine-grained feature context modeling. This paper presents a novel end-to-end re-ranking framework named PIER to tackle the above challenges which still follows the two-stage architecture and contains two mainly modules named FPSM and OCPM. We apply SimHash in FPSM to select top-K candidates from the full permutation based on user's permutation-level interest in an efficient way. Then we design a novel omnidirectional attention mechanism in OCPM to capture the context information in the permutation. Finally, we jointly train these two modules end-to-end by introducing a comparative learning loss. Offline experiment results demonstrate that PIER outperforms baseline models on both public and industrial datasets, and we have successfully deployed PIER on Meituan food delivery platform.
comment: 9 pages, 3 figures
♻ ☆ Reasoning Through Memorization: Nearest Neighbor Knowledge Graph Embeddings
Previous knowledge graph embedding approaches usually map entities to representations and utilize score functions to predict the target entities, yet they typically struggle to reason rare or emerging unseen entities. In this paper, we propose kNN-KGE, a new knowledge graph embedding approach with pre-trained language models, by linearly interpolating its entity distribution with k-nearest neighbors. We compute the nearest neighbors based on the distance in the entity embedding space from the knowledge store. Our approach can allow rare or emerging entities to be memorized explicitly rather than implicitly in model parameters. Experimental results demonstrate that our approach can improve inductive and transductive link prediction results and yield better performance for low-resource settings with only a few triples, which might be easier to reason via explicit memory. Code is available at https://github.com/zjunlp/KNN-KG.
comment: Work in progress
♻ ☆ Two-Stage Constrained Actor-Critic for Short Video Recommendation
The wide popularity of short videos on social media poses new opportunities and challenges to optimize recommender systems on the video-sharing platforms. Users sequentially interact with the system and provide complex and multi-faceted responses, including watch time and various types of interactions with multiple videos. One the one hand, the platforms aims at optimizing the users' cumulative watch time (main goal) in long term, which can be effectively optimized by Reinforcement Learning. On the other hand, the platforms also needs to satisfy the constraint of accommodating the responses of multiple user interactions (auxiliary goals) such like, follow, share etc. In this paper, we formulate the problem of short video recommendation as a Constrained Markov Decision Process (CMDP). We find that traditional constrained reinforcement learning algorithms can not work well in this setting. We propose a novel two-stage constrained actor-critic method: At stage one, we learn individual policies to optimize each auxiliary signal. At stage two, we learn a policy to (i) optimize the main signal and (ii) stay close to policies learned at the first stage, which effectively guarantees the performance of this main policy on the auxiliaries. Through extensive offline evaluations, we demonstrate effectiveness of our method over alternatives in both optimizing the main goal as well as balancing the others. We further show the advantage of our method in live experiments of short video recommendations, where it significantly outperforms other baselines in terms of both watch time and interactions. Our approach has been fully launched in the production system to optimize user experiences on the platform.
comment: arXiv admin note: substantial text overlap with arXiv:2205.13248
♻ ☆ Editing Language Model-based Knowledge Graph Embeddings
Recently decades have witnessed the empirical success of framing Knowledge Graph (KG) embeddings via language models. However, language model-based KG embeddings are usually deployed as static artifacts, which are challenging to modify without re-training after deployment. To address this issue, we propose a new task of editing language model-based KG embeddings in this paper. The proposed task aims to enable data-efficient and fast updates to KG embeddings without damaging the performance of the rest. We build four new datasets: E-FB15k237, A-FB15k237, E-WN18RR, and A-WN18RR, and evaluate several knowledge editing baselines demonstrating the limited ability of previous models to handle the proposed challenging task. We further propose a simple yet strong baseline dubbed KGEditor, which utilizes additional parametric layers of the hyper network to edit/add facts. Comprehensive experimental results demonstrate that KGEditor can perform better when updating specific facts while not affecting the rest with low training resources. Code and datasets will be available in https://github.com/zjunlp/PromptKG/tree/main/deltaKG.
comment: Work in progress and the project website is https://zjunlp.github.io/project/KGE_Editing/
Machine Learning 150
☆ Zero-shot Image-to-Image Translation
Large-scale text-to-image generative models have shown their remarkable ability to synthesize diverse and high-quality images. However, it is still challenging to directly apply these models for editing real images for two reasons. First, it is hard for users to come up with a perfect text prompt that accurately describes every visual detail in the input image. Second, while existing models can introduce desirable changes in certain regions, they often dramatically alter the input content and introduce unexpected changes in unwanted regions. In this work, we propose pix2pix-zero, an image-to-image translation method that can preserve the content of the original image without manual prompting. We first automatically discover editing directions that reflect desired edits in the text embedding space. To preserve the general content structure after editing, we further propose cross-attention guidance, which aims to retain the cross-attention maps of the input image throughout the diffusion process. In addition, our method does not need additional training for these edits and can directly use the existing pre-trained text-to-image diffusion model. We conduct extensive experiments and show that our method outperforms existing and concurrent works for both real and synthetic image editing.
comment: website: https://pix2pixzero.github.io/
☆ Sampling-Based Accuracy Testing of Posterior Estimators for General Inference
Parameter inference, i.e. inferring the posterior distribution of the parameters of a statistical model given some data, is a central problem to many scientific disciplines. Posterior inference with generative models is an alternative to methods such as Markov Chain Monte Carlo, both for likelihood-based and simulation-based inference. However, assessing the accuracy of posteriors encoded in generative models is not straightforward. In this paper, we introduce `distance to random point' (DRP) coverage testing as a method to estimate coverage probabilities of generative posterior estimators. Our method differs from previously-existing coverage-based methods, which require posterior evaluations. We prove that our approach is necessary and sufficient to show that a posterior estimator is optimal. We demonstrate the method on a variety of synthetic examples, and show that DRP can be used to test the results of posterior inference analyses in high-dimensional spaces. We also show that our method can detect non-optimal inferences in cases where existing methods fail.
comment: 15 pages
☆ A Toy Model of Universality: Reverse Engineering How Networks Learn Group Operations
Universality is a key hypothesis in mechanistic interpretability -- that different models learn similar features and circuits when trained on similar tasks. In this work, we study the universality hypothesis by examining how small neural networks learn to implement group composition. We present a novel algorithm by which neural networks may implement composition for any finite group via mathematical representation theory. We then show that networks consistently learn this algorithm by reverse engineering model logits and weights, and confirm our understanding using ablations. By studying networks of differing architectures trained on various groups, we find mixed evidence for universality: using our algorithm, we can completely characterize the family of circuits and features that networks learn on this task, but for a given network the precise circuits learned -- as well as the order they develop -- are arbitrary.
comment: 9 page main body, 1 page references, 12 page appendix
☆ V1T: large-scale mouse V1 response prediction using a Vision Transformer
Accurate predictive models of the visual cortex neural response to natural visual stimuli remain a challenge in computational neuroscience. In this work, we introduce V1T, a novel Vision Transformer based architecture that learns a shared visual and behavioral representation across animals. We evaluate our model on two large datasets recorded from mouse primary visual cortex and outperform previous convolution-based models by more than 12.7% in prediction performance. Moreover, we show that the attention weights learned by the Transformer correlate with the population receptive fields. Our model thus sets a new benchmark for neural response prediction and captures characteristic features of the visual cortex.
☆ RLSbench: Domain Adaptation Under Relaxed Label Shift
Despite the emergence of principled methods for domain adaptation under label shift, the sensitivity of these methods for minor shifts in the class conditional distributions remains precariously under explored. Meanwhile, popular deep domain adaptation heuristics tend to falter when faced with shifts in label proportions. While several papers attempt to adapt these heuristics to accommodate shifts in label proportions, inconsistencies in evaluation criteria, datasets, and baselines, make it hard to assess the state of the art. In this paper, we introduce RLSbench, a large-scale relaxed label shift benchmark, consisting of >500 distribution shift pairs that draw on 14 datasets across vision, tabular, and language modalities and compose them with varying label proportions. First, we evaluate 13 popular domain adaptation methods, demonstrating more widespread failures under label proportion shifts than were previously known. Next, we develop an effective two-step meta-algorithm that is compatible with most deep domain adaptation heuristics: (i) pseudo-balance the data at each epoch; and (ii) adjust the final classifier with (an estimate of) target label distribution. The meta-algorithm improves existing domain adaptation heuristics often by 2--10\% accuracy points under extreme label proportion shifts and has little (i.e., <0.5\%) effect when label proportions do not shift. We hope that these findings and the availability of RLSbench will encourage researchers to rigorously evaluate proposed methods in relaxed label shift settings. Code is publicly available at https://github.com/acmi-lab/RLSbench.
☆ Exploring and Exploiting Decision Boundary Dynamics for Adversarial Robustness ICLR
The robustness of a deep classifier can be characterized by its margins: the decision boundary's distances to natural data points. However, it is unclear whether existing robust training methods effectively increase the margin for each vulnerable point during training. To understand this, we propose a continuous-time framework for quantifying the relative speed of the decision boundary with respect to each individual point. Through visualizing the moving speed of the decision boundary under Adversarial Training, one of the most effective robust training algorithms, a surprising moving-behavior is revealed: the decision boundary moves away from some vulnerable points but simultaneously moves closer to others, decreasing their margins. To alleviate these conflicting dynamics of the decision boundary, we propose Dynamics-aware Robust Training (DyART), which encourages the decision boundary to engage in movement that prioritizes increasing smaller margins. In contrast to prior works, DyART directly operates on the margins rather than their indirect approximations, allowing for more targeted and effective robustness improvement. Experiments on the CIFAR-10 and Tiny-ImageNet datasets verify that DyART alleviates the conflicting dynamics of the decision boundary and obtains improved robustness under various perturbation sizes compared to the state-of-the-art defenses. Our code is available at https://github.com/Yuancheng-Xu/Dynamics-Aware-Robust-Training.
comment: Published at International Conference on Learning Representations (ICLR) 2023
☆ Detection and Localization of Melanoma Skin Cancer in Histopathological Whole Slide Images
Melanoma diagnosed and treated in its early stages can increase the survival rate. A projected increase in skin cancer incidents and a dearth of dermatopathologists have emphasized the need for computational pathology (CPATH) systems. CPATH systems with deep learning (DL) models have the potential to identify the presence of melanoma by exploiting underlying morphological and cellular features. This paper proposes a DL method to detect melanoma and distinguish between normal skin and benign/malignant melanocytic lesions in Whole Slide Images (WSI). Our method detects lesions with high accuracy and localizes them on a WSI to identify potential regions of interest for pathologists. Interestingly, our DL method relies on using a single CNN network to create localization maps first and use them to perform slide-level predictions to determine patients who have melanoma. Our best model provides favorable patch-wise classification results with a 0.992 F1 score and 0.99 sensitivity on unseen data.
comment: Submitted to EUSIPCO 23
☆ LAVA: Granular Neuron-Level Explainable AI for Alzheimer's Disease Assessment from Fundus Images
Alzheimer's Disease (AD) is a progressive neurodegenerative disease and the leading cause of dementia. Early diagnosis is critical for patients to benefit from potential intervention and treatment. The retina has been hypothesized as a diagnostic site for AD detection owing to its anatomical connection with the brain. Developed AI models for this purpose have yet to provide a rational explanation about the decision and neither infer the stage of disease's progression. Along this direction, we propose a novel model-agnostic explainable-AI framework, called Granular Neuron-level Explainer (LAVA), an interpretation prototype that probes into intermediate layers of the Convolutional Neural Network (CNN) models to assess the AD continuum directly from the retinal imaging without longitudinal or clinical evaluation. This method is applied to validate the retinal vasculature as a biomarker and diagnostic modality for Alzheimer's Disease (AD) evaluation. UK Biobank cognitive tests and vascular morphological features suggest LAVA shows strong promise and effectiveness in identifying AD stages across the progression continuum.
comment: 27 pages, 11 figures
☆ Neural Collapse Inspired Feature-Classifier Alignment for Few-Shot Class Incremental Learning ICLR 2023
Few-shot class-incremental learning (FSCIL) has been a challenging problem as only a few training samples are accessible for each novel class in the new sessions. Finetuning the backbone or adjusting the classifier prototypes trained in the prior sessions would inevitably cause a misalignment between the feature and classifier of old classes, which explains the well-known catastrophic forgetting problem. In this paper, we deal with this misalignment dilemma in FSCIL inspired by the recently discovered phenomenon named neural collapse, which reveals that the last-layer features of the same class will collapse into a vertex, and the vertices of all classes are aligned with the classifier prototypes, which are formed as a simplex equiangular tight frame (ETF). It corresponds to an optimal geometric structure for classification due to the maximized Fisher Discriminant Ratio. We propose a neural collapse inspired framework for FSCIL. A group of classifier prototypes are pre-assigned as a simplex ETF for the whole label space, including the base session and all the incremental sessions. During training, the classifier prototypes are not learnable, and we adopt a novel loss function that drives the features into their corresponding prototypes. Theoretical analysis shows that our method holds the neural collapse optimality and does not break the feature-classifier alignment in an incremental fashion. Experiments on the miniImageNet, CUB-200, and CIFAR-100 datasets demonstrate that our proposed framework outperforms the state-of-the-art performances. Code address: https://github.com/NeuralCollapseApplications/FSCIL
comment: ICLR 2023 (Notable-top-25%)
☆ Asymptotically Minimax Optimal Fixed-Budget Best Arm Identification for Expected Simple Regret Minimization
We investigate fixed-budget best arm identification (BAI) for expected simple regret minimization. In each round of an adaptive experiment, a decision maker draws one of multiple treatment arms based on past observations and subsequently observes the outcomes of the chosen arm. After the experiment, the decision maker recommends a treatment arm with the highest projected outcome. We evaluate this decision in terms of the expected simple regret, a difference between the expected outcomes of the best and recommended treatment arms. Due to the inherent uncertainty, we evaluate the regret using the minimax criterion. For distributions with fixed variances (location-shift models), such as Gaussian distributions, we derive asymptotic lower bounds for the worst-case expected simple regret. Then, we show that the Random Sampling (RS)-Augmented Inverse Probability Weighting (AIPW) strategy proposed by Kato et al. (2022) is asymptotically minimax optimal in the sense that the leading factor of its worst-case expected simple regret asymptotically matches our derived worst-case lower bound. Our result indicates that, for location-shift models, the optimal RS-AIPW strategy draws treatment arms with varying probabilities based on their variances. This result contrasts with the results of Bubeck et al. (2011), which shows that drawing each treatment arm with an equal ratio is minimax optimal in a bounded outcome setting.
☆ Robust Subtask Learning for Compositional Generalization
Compositional reinforcement learning is a promising approach for training policies to perform complex long-horizon tasks. Typically, a high-level task is decomposed into a sequence of subtasks and a separate policy is trained to perform each subtask. In this paper, we focus on the problem of training subtask policies in a way that they can be used to perform any task; here, a task is given by a sequence of subtasks. We aim to maximize the worst-case performance over all tasks as opposed to the average-case performance. We formulate the problem as a two agent zero-sum game in which the adversary picks the sequence of subtasks. We propose two RL algorithms to solve this game: one is an adaptation of existing multi-agent RL algorithms to our setting and the other is an asynchronous version which enables parallel training of subtask policies. We evaluate our approach on two multi-task environments with continuous states and actions and demonstrate that our algorithms outperform state-of-the-art baselines.
☆ Learning disentangled representations for explainable chest X-ray classification using Dirichlet VAEs SP
This study explores the use of the Dirichlet Variational Autoencoder (DirVAE) for learning disentangled latent representations of chest X-ray (CXR) images. Our working hypothesis is that distributional sparsity, as facilitated by the Dirichlet prior, will encourage disentangled feature learning for the complex task of multi-label classification of CXR images. The DirVAE is trained using CXR images from the CheXpert database, and the predictive capacity of multi-modal latent representations learned by DirVAE models is investigated through implementation of an auxiliary multi-label classification task, with a view to enforce separation of latent factors according to class-specific features. The predictive performance and explainability of the latent space learned using the DirVAE were quantitatively and qualitatively assessed, respectively, and compared with a standard Gaussian prior-VAE (GVAE). We introduce a new approach for explainable multi-label classification in which we conduct gradient-guided latent traversals for each class of interest. Study findings indicate that the DirVAE is able to disentangle latent factors into class-specific visual features, a property not afforded by the GVAE, and achieve a marginal increase in predictive performance relative to GVAE. We generate visual examples to show that our explainability method, when applied to the trained DirVAE, is able to highlight regions in CXR images that are clinically relevant to the class(es) of interest and additionally, can identify cases where classification relies on spurious feature correlations.
comment: 13 pages, 8 figures, to be published in SPIE Medical Imaging 2023
☆ MuG: A Multimodal Classification Benchmark on Game Data with Tabular, Textual, and Visual Fields
Multimodal learning has attracted the interest of the machine learning community due to its great potential in a variety of applications. To help achieve this potential, we propose a multimodal benchmark MuG with eight datasets allowing researchers to test the multimodal perceptron capabilities of their models. These datasets are collected from four different genres of games that cover tabular, textual, and visual modalities. We conduct multi-aspect data analysis to provide insights into the benchmark, including label balance ratios, percentages of missing features, distributions of data within each modality, and the correlations between labels and input modalities. We further present experimental results obtained by several state-of-the-art unimodal classifiers and multimodal classifiers, which demonstrate the challenging and multimodal-dependent properties of the benchmark. MuG is released at https://github.com/lujiaying/MUG-Bench with the data, documents, tutorials, and implemented baselines. Extensions of MuG are welcomed to facilitate the progress of research in multimodal learning problems.
☆ Concrete Safety for ML Problems: System Safety for ML Development and Assessment
Many stakeholders struggle to make reliances on ML-driven systems due to the risk of harm these systems may cause. Concerns of trustworthiness, unintended social harms, and unacceptable social and ethical violations undermine the promise of ML advancements. Moreover, such risks in complex ML-driven systems present a special challenge as they are often difficult to foresee, arising over periods of time, across populations, and at scale. These risks often arise not from poor ML development decisions or low performance directly but rather emerge through the interactions amongst ML development choices, the context of model use, environmental factors, and the effects of a model on its target. Systems safety engineering is an established discipline with a proven track record of identifying and managing risks even in high-complexity sociotechnical systems. In this work, we apply a state-of-the-art systems safety approach to concrete applications of ML with notable social and ethical risks to demonstrate a systematic means for meeting the assurance requirements needed to argue for safe and trustworthy ML in sociotechnical systems.
comment: arXiv admin note: text overlap with arXiv:2211.04602
☆ U-Clip: On-Average Unbiased Stochastic Gradient Clipping
U-Clip is a simple amendment to gradient clipping that can be applied to any iterative gradient optimization algorithm. Like regular clipping, U-Clip involves using gradients that are clipped to a prescribed size (e.g. with component wise or norm based clipping) but instead of discarding the clipped portion of the gradient, U-Clip maintains a buffer of these values that is added to the gradients on the next iteration (before clipping). We show that the cumulative bias of the U-Clip updates is bounded by a constant. This implies that the clipped updates are unbiased on average. Convergence follows via a lemma that guarantees convergence with updates $u_i$ as long as $\sum_{i=1}^t (u_i - g_i) = o(t)$ where $g_i$ are the gradients. Extensive experimental exploration is performed on CIFAR10 with further validation given on ImageNet.
☆ Adaptive Parameterization of Deep Learning Models for Federated Learning
Federated Learning offers a way to train deep neural networks in a distributed fashion. While this addresses limitations related to distributed data, it incurs a communication overhead as the model parameters or gradients need to be exchanged regularly during training. This can be an issue with large scale distribution of learning asks and negate the benefit of the respective resource distribution. In this paper, we we propose to utilise parallel Adapters for Federated Learning. Using various datasets, we show that Adapters can be applied with different Federated Learning techniques. We highlight that our approach can achieve similar inference performance compared to training the full model while reducing the communication overhead drastically. We further explore the applicability of Adapters in cross-silo and cross-device settings, as well as different non-IID data distributions.
☆ GPS++: Reviving the Art of Message Passing for Molecular Property Prediction
We present GPS++, a hybrid Message Passing Neural Network / Graph Transformer model for molecular property prediction. Our model integrates a well-tuned local message passing component and biased global attention with other key ideas from prior literature to achieve state-of-the-art results on large-scale molecular dataset PCQM4Mv2. Through a thorough ablation study we highlight the impact of individual components and, contrary to expectations set by recent trends, find that nearly all of the model's performance can be maintained without any use of global self-attention. We also show that our approach is significantly more accurate than prior art when 3D positional information is not available.
comment: arXiv admin note: text overlap with arXiv:2212.02229
☆ Efficient Online Reinforcement Learning with Offline Data
Sample efficiency and exploration remain major challenges in online reinforcement learning (RL). A powerful approach that can be applied to address these issues is the inclusion of offline data, such as prior trajectories from a human expert or a sub-optimal exploration policy. Previous methods have relied on extensive modifications and additional complexity to ensure the effective use of this data. Instead, we ask: can we simply apply existing off-policy methods to leverage offline data when learning online? In this work, we demonstrate that the answer is yes; however, a set of minimal but important changes to existing off-policy RL algorithms are required to achieve reliable performance. We extensively ablate these design choices, demonstrating the key factors that most affect performance, and arrive at a set of recommendations that practitioners can readily apply, whether their data comprise a small number of expert demonstrations or large volumes of sub-optimal trajectories. We see that correct application of these simple recommendations can provide a $\mathbf{2.5\times}$ improvement over existing approaches across a diverse set of competitive benchmarks, with no additional computational overhead.
☆ On Over-Squashing in Message Passing Neural Networks: The Impact of Width, Depth, and Topology
Message Passing Neural Networks (MPNNs) are instances of Graph Neural Networks that leverage the graph to send messages over the edges. This inductive bias leads to a phenomenon known as over-squashing, where a node feature is insensitive to information contained at distant nodes. Despite recent methods introduced to mitigate this issue, an understanding of the causes for over-squashing and of possible solutions are lacking. In this theoretical work, we prove that: (i) Neural network width can mitigate over-squashing, but at the cost of making the whole network more sensitive; (ii) Conversely, depth cannot help mitigate over-squashing: increasing the number of layers leads to over-squashing being dominated by vanishing gradients; (iii) The graph topology plays the greatest role, since over-squashing occurs between nodes at high commute (access) time. Our analysis provides a unified framework to study different recent methods introduced to cope with over-squashing and serves as a justification for a class of methods that fall under `graph rewiring'.
comment: 24 pages
☆ Private GANs, Revisited
We show that the canonical approach for training differentially private GANs -- updating the discriminator with differentially private stochastic gradient descent (DPSGD) -- can yield significantly improved results after modifications to training. Existing instantiations of this approach neglect to consider how adding noise only to discriminator updates disrupts the careful balance between the generator and discriminator necessary for successful GAN training. We show that a simple fix -- taking more discriminator steps between generator steps -- restores parity and improves results. Additionally, with the goal of restoring parity between the generator and discriminator, we experiment with other modifications to improve discriminator training and see further improvements in generation quality. Our results demonstrate that on standard benchmarks, DPSGD outperforms all alternative GAN privatization schemes.
☆ Bitrate-Constrained DRO: Beyond Worst Case Robustness To Unknown Group Shifts
Training machine learning models robust to distribution shifts is critical for real-world applications. Some robust training algorithms (e.g., Group DRO) specialize to group shifts and require group information on all training points. Other methods (e.g., CVaR DRO) that do not need group annotations can be overly conservative, since they naively upweight high loss points which may form a contrived set that does not correspond to any meaningful group in the real world (e.g., when the high loss points are randomly mislabeled training points). In this work, we address limitations in prior approaches by assuming a more nuanced form of group shift: conditioned on the label, we assume that the true group function (indicator over group) is simple. For example, we may expect that group shifts occur along low bitrate features (e.g., image background, lighting). Thus, we aim to learn a model that maintains high accuracy on simple group functions realized by these low bitrate features, that need not spend valuable model capacity achieving high accuracy on contrived groups of examples. Based on this, we consider the two-player game formulation of DRO where the adversary's capacity is bitrate-constrained. Our resulting practical algorithm, Bitrate-Constrained DRO (BR-DRO), does not require group information on training samples yet matches the performance of Group DRO on datasets that have training group annotations and that of CVaR DRO on long-tailed distributions. Our theoretical analysis reveals that in some settings BR-DRO objective can provably yield statistically efficient and less conservative solutions than unconstrained CVaR DRO.
☆ Curriculum Graph Machine Learning: A Survey
Graph machine learning has been extensively studied in both academia and industry. However, in the literature, most existing graph machine learning models are designed to conduct training with data samples in a random order, which may suffer from suboptimal performance due to ignoring the importance of different graph data samples and their training orders for the model optimization status. To tackle this critical problem, curriculum graph machine learning (Graph CL), which integrates the strength of graph machine learning and curriculum learning, arises and attracts an increasing amount of attention from the research community. Therefore, in this paper, we comprehensively overview approaches on Graph CL and present a detailed survey of recent advances in this direction. Specifically, we first discuss the key challenges of Graph CL and provide its formal problem definition. Then, we categorize and summarize existing methods into three classes based on three kinds of graph machine learning tasks, i.e., node-level, link-level, and graph-level tasks. Finally, we share our thoughts on future research directions. To the best of our knowledge, this paper is the first survey for curriculum graph machine learning.
☆ Dropout Injection at Test Time for Post Hoc Uncertainty Quantification in Neural Networks
Among Bayesian methods, Monte-Carlo dropout provides principled tools for evaluating the epistemic uncertainty of neural networks. Its popularity recently led to seminal works that proposed activating the dropout layers only during inference for evaluating uncertainty. This approach, which we call dropout injection, provides clear benefits over its traditional counterpart (which we call embedded dropout) since it allows one to obtain a post hoc uncertainty measure for any existing network previously trained without dropout, avoiding an additional, time-consuming training process. Unfortunately, no previous work compared injected and embedded dropout; therefore, we provide the first thorough investigation, focusing on regression problems. The main contribution of our work is to provide guidelines on the effective use of injected dropout so that it can be a practical alternative to the current use of embedded dropout. In particular, we show that its effectiveness strongly relies on a suitable scaling of the corresponding uncertainty measure, and we discuss the trade-off between negative log-likelihood and calibration error as a function of the scale factor. Experimental results on UCI data sets and crowd counting benchmarks support our claim that dropout injection can effectively behave as a competitive post hoc uncertainty quantification technique.
☆ In Search of Insights, Not Magic Bullets: Towards Demystification of the Model Selection Dilemma in Heterogeneous Treatment Effect Estimation
Personalized treatment effect estimates are often of interest in high-stakes applications -- thus, before deploying a model estimating such effects in practice, one needs to be sure that the best candidate from the ever-growing machine learning toolbox for this task was chosen. Unfortunately, due to the absence of counterfactual information in practice, it is usually not possible to rely on standard validation metrics for doing so, leading to a well-known model selection dilemma in the treatment effect estimation literature. While some solutions have recently been investigated, systematic understanding of the strengths and weaknesses of different model selection criteria is still lacking. In this paper, instead of attempting to declare a global `winner', we therefore empirically investigate success- and failure modes of different selection criteria. We highlight that there is a complex interplay between selection strategies, candidate estimators and the DGP used for testing, and provide interesting insights into the relative (dis)advantages of different criteria alongside desiderata for the design of further illuminating empirical studies in this context.
☆ Joint Edge-Model Sparse Learning is Provably Efficient for Graph Neural Networks
Due to the significant computational challenge of training large-scale graph neural networks (GNNs), various sparse learning techniques have been exploited to reduce memory and storage costs. Examples include \textit{graph sparsification} that samples a subgraph to reduce the amount of data aggregation and \textit{model sparsification} that prunes the neural network to reduce the number of trainable weights. Despite the empirical successes in reducing the training cost while maintaining the test accuracy, the theoretical generalization analysis of sparse learning for GNNs remains elusive. To the best of our knowledge, this paper provides the first theoretical characterization of joint edge-model sparse learning from the perspective of sample complexity and convergence rate in achieving zero generalization error. It proves analytically that both sampling important nodes and pruning neurons with the lowest-magnitude can reduce the sample complexity and improve convergence without compromising the test accuracy. Although the analysis is centered on two-layer GNNs with structural constraints on data, the insights are applicable to more general setups and justified by both synthetic and practical citation datasets.
☆ Holistic Deep-Reinforcement-Learning-based Training of Autonomous Navigation Systems
In recent years, Deep Reinforcement Learning emerged as a promising approach for autonomous navigation of ground vehicles and has been utilized in various areas of navigation such as cruise control, lane changing, or obstacle avoidance. However, most research works either focus on providing an end-to-end solution training the whole system using Deep Reinforcement Learning or focus on one specific aspect such as local motion planning. This however, comes along with a number of problems such as catastrophic forgetfulness, inefficient navigation behavior, and non-optimal synchronization between different entities of the navigation stack. In this paper, we propose a holistic Deep Reinforcement Learning training approach in which the training procedure is involving all entities of the navigation stack. This should enhance the synchronization between- and understanding of all entities of the navigation stack and as a result, improve navigational performance. We trained several agents with a number of different observation spaces to study the impact of different input on the navigation behavior of the agent. In profound evaluations against multiple learning-based and classic model-based navigation approaches, our proposed agent could outperform the baselines in terms of efficiency and safety attaining shorter path lengths, less roundabout paths, and less collisions.
comment: 8 pages, 6 figures
☆ Energy-based Out-of-Distribution Detection for Graph Neural Networks ICLR 2023
Learning on graphs, where instance nodes are inter-connected, has become one of the central problems for deep learning, as relational structures are pervasive and induce data inter-dependence which hinders trivial adaptation of existing approaches that assume inputs to be i.i.d.~sampled. However, current models mostly focus on improving testing performance of in-distribution data and largely ignore the potential risk w.r.t. out-of-distribution (OOD) testing samples that may cause negative outcome if the prediction is overconfident on them. In this paper, we investigate the under-explored problem, OOD detection on graph-structured data, and identify a provably effective OOD discriminator based on an energy function directly extracted from graph neural networks trained with standard classification loss. This paves a way for a simple, powerful and efficient OOD detection model for GNN-based learning on graphs, which we call GNNSafe. It also has nice theoretical properties that guarantee an overall distinguishable margin between the detection scores for in-distribution and OOD samples, which, more critically, can be further strengthened by a learning-free energy belief propagation scheme. For comprehensive evaluation, we introduce new benchmark settings that evaluate the model for detecting OOD data from both synthetic and real distribution shifts (cross-domain graph shifts and temporal graph shifts). The results show that GNNSafe achieves up to $17.0\%$ AUROC improvement over state-of-the-arts and it could serve as simple yet strong baselines in such an under-developed area.
comment: Accepted by International Conference on Learning Representations (ICLR 2023)
☆ Beyond Statistical Similarity: Rethinking Metrics for Deep Generative Models in Engineering Design
Deep generative models, such as Variational Autoencoders (VAEs), Generative Adversarial Networks (GANs), Diffusion Models, and Transformers, have shown great promise in a variety of applications, including image and speech synthesis, natural language processing, and drug discovery. However, when applied to engineering design problems, evaluating the performance of these models can be challenging, as traditional statistical metrics based on likelihood may not fully capture the requirements of engineering applications. This paper doubles as a review and a practical guide to evaluation metrics for deep generative models (DGMs) in engineering design. We first summarize well-accepted `classic' evaluation metrics for deep generative models grounded in machine learning theory and typical computer science applications. Using case studies, we then highlight why these metrics seldom translate well to design problems but see frequent use due to the lack of established alternatives. Next, we curate a set of design-specific metrics which have been proposed across different research communities and can be used for evaluating deep generative models. These metrics focus on unique requirements in design and engineering, such as constraint satisfaction, functional performance, novelty, and conditioning. We structure our review and discussion as a set of practical selection criteria and usage guidelines. Throughout our discussion, we apply the metrics to models trained on simple 2-dimensional example problems. Finally, to illustrate the selection process and classic usage of the presented metrics, we evaluate three deep generative models on a multifaceted bicycle frame design problem considering performance target achievement, design novelty, and geometric constraints. We publicly release the code for the datasets, models, and metrics used throughout the paper at decode.mit.edu/projects/metrics/.
☆ An Empirical Analysis of Fairness Notions under Differential Privacy AAAI
Recent works have shown that selecting an optimal model architecture suited to the differential privacy setting is necessary to achieve the best possible utility for a given privacy budget using differentially private stochastic gradient descent (DP-SGD)(Tramer and Boneh 2020; Cheng et al. 2022). In light of these findings, we empirically analyse how different fairness notions, belonging to distinct classes of statistical fairness criteria (independence, separation and sufficiency), are impacted when one selects a model architecture suitable for DP-SGD, optimized for utility. Using standard datasets from ML fairness literature, we show using a rigorous experimental protocol, that by selecting the optimal model architecture for DP-SGD, the differences across groups concerning the relevant fairness metrics (demographic parity, equalized odds and predictive parity) more often decrease or are negligibly impacted, compared to the non-private baseline, for which optimal model architecture has also been selected to maximize utility. These findings challenge the understanding that differential privacy will necessarily exacerbate unfairness in deep learning models trained on biased datasets.
comment: Accepted for oral presentation at the The Fourth AAAI Workshop on Privacy-Preserving Artificial Intelligence (PPAI-23) https://aaai-ppai23.github.io/#accepted_papers
☆ Spectral Augmentations for Graph Contrastive Learning AISTATS 2023
Contrastive learning has emerged as a premier method for learning representations with or without supervision. Recent studies have shown its utility in graph representation learning for pre-training. Despite successes, the understanding of how to design effective graph augmentations that can capture structural properties common to many different types of downstream graphs remains incomplete. We propose a set of well-motivated graph transformation operations derived via graph spectral analysis to provide a bank of candidates when constructing augmentations for a graph contrastive objective, enabling contrastive learning to capture useful structural representation from pre-training graph datasets. We first present a spectral graph cropping augmentation that involves filtering nodes by applying thresholds to the eigenvalues of the leading Laplacian eigenvectors. Our second novel augmentation reorders the graph frequency components in a structural Laplacian-derived position graph embedding. Further, we introduce a method that leads to improved views of local subgraphs by performing alignment via global random walk embeddings. Our experimental results indicate consistent improvements in out-of-domain graph data transfer compared to state-of-the-art graph contrastive learning methods, shedding light on how to design a graph learner that is able to learn structural properties common to diverse graph types.
comment: To appear in AISTATS 2023
☆ GAT: Guided Adversarial Training with Pareto-optimal Auxiliary Tasks
While leveraging additional training data is well established to improve adversarial robustness, it incurs the unavoidable cost of data collection and the heavy computation to train models. To mitigate the costs, we propose \textit{Guided Adversarial Training } (GAT), a novel adversarial training technique that exploits auxiliary tasks under a limited set of training data. Our approach extends single-task models into multi-task models during the min-max optimization of adversarial training, and drives the loss optimization with a regularization of the gradient curvature across multiple tasks. GAT leverages two types of auxiliary tasks: self-supervised tasks, where the labels are generated automatically, and domain-knowledge tasks, where human experts provide additional labels. Experimentally, under limited data, GAT increases the robust accuracy on CIFAR-10 up to four times (from 11% to 42% robust accuracy) and the robust AUC of CheXpert medical imaging dataset from 50\% to 83\%. On the full CIFAR-10 dataset, GAT outperforms eight state-of-the-art adversarial training strategies. Our large study across five datasets and six tasks demonstrates that task augmentation is an efficient alternative to data augmentation, and can be key to achieving both clean and robust performances.
☆ Rethinking Gauss-Newton for learning over-parameterized models
Compared to gradient descent, Gauss-Newton's method (GN) and variants are known to converge faster to local optima at the expense of a higher computational cost per iteration. Still, GN is not widely used for optimizing deep neural networks despite a constant effort to reduce their higher computational cost. In this work, we propose to take a step back and re-think the properties of GN in light of recent advances in the dynamics of gradient flows of over-parameterized models and the implicit bias they induce. We first prove a fast global convergence result for the continuous-time limit of the generalized GN in the over-parameterized regime. We then show empirically that GN exhibits both a kernel regime where it generalizes as well as gradient flows, and a feature learning regime where GN induces an implicit bias for selecting global solutions that systematically under-performs those found by a gradient flow. Importantly, we observed this phenomenon even with enough computational budget to perform exact GN steps over the total training objective. This study suggests the need to go beyond improving the computational cost of GN for over-parametrized models towards designing new methods that can trade off optimization speed and the quality of their implicit bias.
☆ Controllable Lexical Simplification for English
Fine-tuning Transformer-based approaches have recently shown exciting results on sentence simplification task. However, so far, no research has applied similar approaches to the Lexical Simplification (LS) task. In this paper, we present ConLS, a Controllable Lexical Simplification system fine-tuned with T5 (a Transformer-based model pre-trained with a BERT-style approach and several other tasks). The evaluation results on three datasets (LexMTurk, BenchLS, and NNSeval) have shown that our model performs comparable to LSBert (the current state-of-the-art) and even outperforms it in some cases. We also conducted a detailed comparison on the effectiveness of control tokens to give a clear view of how each token contributes to the model.
☆ Arena-Web -- A Web-based Development and Benchmarking Platform for Autonomous Navigation Approaches
In recent years, mobile robot navigation approaches have become increasingly important due to various application areas ranging from healthcare to warehouse logistics. In particular, Deep Reinforcement Learning approaches have gained popularity for robot navigation but are not easily accessible to non-experts and complex to develop. In recent years, efforts have been made to make these sophisticated approaches accessible to a wider audience. In this paper, we present Arena-Web, a web-based development and evaluation suite for developing, training, and testing DRL-based navigation planners for various robotic platforms and scenarios. The interface is designed to be intuitive and engaging to appeal to non-experts and make the technology accessible to a wider audience. With Arena-Web and its interface, training and developing Deep Reinforcement Learning agents is simplified and made easy without a single line of code. The web-app is free to use and openly available under the link stated in the supplementary materials.
comment: 10 pages, 9 figures
☆ Label Assisted Autoencoder for Anomaly Detection in Power Generation Plants
One of the critical factors that drive the economic development of a country and guarantee the sustainability of its industries is the constant availability of electricity. This is usually provided by the national electric grid. However, in developing countries where companies are emerging on a constant basis including telecommunication industries, those are still experiencing a non-stable electricity supply. Therefore, they have to rely on generators to guarantee their full functionality. Those generators depend on fuel to function and the rate of consumption gets usually high, if not monitored properly. Monitoring operation is usually carried out by a (non-expert) human. In some cases, this could be a tedious process, as some companies have reported an exaggerated high consumption rate. This work proposes a label assisted autoencoder for anomaly detection in the fuel consumed by power generating plants. In addition to the autoencoder model, we added a labelling assistance module that checks if an observation is labelled, the label is used to check the veracity of the corresponding anomaly classification given a threshold. A consensus is then reached on whether training should stop or whether the threshold should be updated or the training should continue with the search for hyper-parameters. Results show that the proposed model is highly efficient for reading anomalies with a detection accuracy of $97.20\%$ which outperforms the existing model of $96.1\%$ accuracy trained on the same dataset. In addition, the proposed model is able to classify the anomalies according to their degree of severity.
comment: Submitted to Journal
☆ Findings of the TSAR-2022 Shared Task on Multilingual Lexical Simplification
We report findings of the TSAR-2022 shared task on multilingual lexical simplification, organized as part of the Workshop on Text Simplification, Accessibility, and Readability TSAR-2022 held in conjunction with EMNLP 2022. The task called the Natural Language Processing research community to contribute with methods to advance the state of the art in multilingual lexical simplification for English, Portuguese, and Spanish. A total of 14 teams submitted the results of their lexical simplification systems for the provided test data. Results of the shared task indicate new benchmarks in Lexical Simplification with English lexical simplification quantitative results noticeably higher than those obtained for Spanish and (Brazilian) Portuguese.
☆ Intra-operative Brain Tumor Detection with Deep Learning-Optimized Hyperspectral Imaging SP
Surgery for gliomas (intrinsic brain tumors), especially when low-grade, is challenging due to the infiltrative nature of the lesion. Currently, no real-time, intra-operative, label-free and wide-field tool is available to assist and guide the surgeon to find the relevant demarcations for these tumors. While marker-based methods exist for the high-grade glioma case, there is no convenient solution available for the low-grade case; thus, marker-free optical techniques represent an attractive option. Although RGB imaging is a standard tool in surgical microscopes, it does not contain sufficient information for tissue differentiation. We leverage the richer information from hyperspectral imaging (HSI), acquired with a snapscan camera in the 468-787 nm range, coupled to a surgical microscope, to build a deep-learning-based diagnostic tool for cancer resection with potential for intra-operative guidance. However, the main limitation of the HSI snapscan camera is the image acquisition time, limiting its widespread deployment in the operation theater. Here, we investigate the effect of HSI channel reduction and pre-selection to scope the design space for the development of cheaper and faster sensors. Neural networks are used to identify the most important spectral channels for tumor tissue differentiation, optimizing the trade-off between the number of channels and precision to enable real-time intra-surgical application. We evaluate the performance of our method on a clinical dataset that was acquired during surgery on five patients. By demonstrating the possibility to efficiently detect low-grade glioma, these results can lead to better cancer resection demarcations, potentially improving treatment effectiveness and patient outcome.
comment: SPIE Photonics West 2023 conference Optical Biopsy XXI: Toward Real-Time Spectroscopic Imaging and Diagnosis. 18 pages, 11 figures
☆ Variational Information Pursuit for Interpretable Predictions
There is a growing interest in the machine learning community in developing predictive algorithms that are "interpretable by design". Towards this end, recent work proposes to make interpretable decisions by sequentially asking interpretable queries about data until a prediction can be made with high confidence based on the answers obtained (the history). To promote short query-answer chains, a greedy procedure called Information Pursuit (IP) is used, which adaptively chooses queries in order of information gain. Generative models are employed to learn the distribution of query-answers and labels, which is in turn used to estimate the most informative query. However, learning and inference with a full generative model of the data is often intractable for complex tasks. In this work, we propose Variational Information Pursuit (V-IP), a variational characterization of IP which bypasses the need for learning generative models. V-IP is based on finding a query selection strategy and a classifier that minimizes the expected cross-entropy between true and predicted labels. We then demonstrate that the IP strategy is the optimal solution to this problem. Therefore, instead of learning generative models, we can use our optimal strategy to directly pick the most informative query given any history. We then develop a practical algorithm by defining a finite-dimensional parameterization of our strategy and classifier using deep networks and train them end-to-end using our objective. Empirically, V-IP is 10-100x faster than IP on different Vision and NLP tasks with competitive performance. Moreover, V-IP finds much shorter query chains when compared to reinforcement learning which is typically used in sequential-decision-making problems. Finally, we demonstrate the utility of V-IP on challenging tasks like medical diagnosis where the performance is far superior to the generative modelling approach.
☆ Probabilistic Contrastive Learning Recovers the Correct Aleatoric Uncertainty of Ambiguous Inputs
Contrastively trained encoders have recently been proven to invert the data-generating process: they encode each input, e.g., an image, into the true latent vector that generated the image (Zimmermann et al., 2021). However, real-world observations often have inherent ambiguities. For instance, images may be blurred or only show a 2D view of a 3D object, so multiple latents could have generated them. This makes the true posterior for the latent vector probabilistic with heteroscedastic uncertainty. In this setup, we extend the common InfoNCE objective and encoders to predict latent distributions instead of points. We prove that these distributions recover the correct posteriors of the data-generating process, including its level of aleatoric uncertainty, up to a rotation of the latent space. In addition to providing calibrated uncertainty estimates, these posteriors allow the computation of credible intervals in image retrieval. They comprise images with the same latent as a given query, subject to its uncertainty.
☆ Solving Maxwell's Equation in 2D with Neural Networks with Local Converging Inputs
In this paper we apply neural networks with local converging inputs (NNLCI), originally introduced in [arXiv:2109.09316], to solve the two dimensional Maxwell's equation around perfect electric conductors (PECs). The input to the networks consist of local patches of low cost numerical solutions to the equation computed on two coarse grids, and the output is a more accurate solution at the center of the local patch. We apply the recently developed second order finite difference method [arXiv:2209.00740] to generate the input and training data which captures the scattering of electromagnetic waves off of a PEC at a given terminal time. The advantage of NNLCI is that once trained it offers an efficient alternative to costly high-resolution conventional numerical methods; our numerical experiments indicate the computational complexity saving by a factor of $8^3$ in terms of the number of spatial-temporal grid points. In contrast with existing research work on applying neural networks to directly solve PDEs, our method takes advantage of the local domain of dependence of the Maxwell's equation in the input solution patches, and is therefore simpler, yet still robust. We demonstrate that we can train our neural network on some PECs to predict accurate solutions to different PECs with quite different geometries from any of the training examples.
☆ NA-SODINN: a deep learning algorithm for exoplanet image detection based on residual noise regimes
Supervised machine learning was recently introduced in high-contrast imaging (HCI) through the SODINN algorithm, a convolutional neural network designed for exoplanet detection in angular differential imaging (ADI) data sets. The benchmarking of HCI algorithms within the Exoplanet Imaging Data Challenge (EIDC) showed that (i) SODINN can produce a high number of false positives in the final detection maps, and (ii) algorithms processing images in a more local manner perform better. This work aims to improve the SODINN detection performance by introducing new local processing approaches and adapting its learning process accordingly. We propose NA-SODINN, a new deep learning architecture that better captures image noise correlations by training an independent SODINN model per noise regime over the processed frame. The identification of these noise regimes is based on a novel technique, named PCA-pmaps, which allows to estimate the distance from the star in the image from which background noise starts to dominate over residual speckle noise. NA-SODINN is also fed with local discriminators, such as S/N curves, which complement spatio-temporal feature maps when training the model.Our new approach is tested against its predecessor, as well as two SODINN-based hybrid models and a more standard annular-PCA approach, through local ROC analysis of ADI sequences from VLT/SPHERE and Keck/NIRC-2 instruments. Results show that NA-SODINN enhances SODINN in both the sensitivity and specificity, especially in the speckle-dominated noise regime. NA-SODINN is also benchmarked against the complete set of submitted detection algorithms in EIDC, in which we show that its final detection score matches or outperforms the most powerful detection algorithms, reaching a performance similar to that of the Regime Switching Model algorithm.
comment: Submitted to A&A journal.Comments on the submitted version are welcome
☆ An Unsupervised Framework for Joint MRI Super Resolution and Gibbs Artifact Removal
The k-space data generated from magnetic resonance imaging (MRI) is only a finite sampling of underlying signals. Therefore, MRI images often suffer from low spatial resolution and Gibbs ringing artifacts. Previous studies tackled these two problems separately, where super resolution methods tend to enhance Gibbs artifacts, whereas Gibbs ringing removal methods tend to blur the images. It is also a challenge that high resolution ground truth is hard to obtain in clinical MRI. In this paper, we propose an unsupervised learning framework for both MRI super resolution and Gibbs artifacts removal without using high resolution ground truth. Furthermore, we propose regularization methods to improve the model's generalizability across out-of-distribution MRI images. We evaluated our proposed methods with other state-of-the-art methods on eight MRI datasets with various contrasts and anatomical structures. Our method not only achieves the best SR performance but also significantly reduces the Gibbs artifacts. Our method also demonstrates good generalizability across different datasets, which is beneficial to clinical applications where training data are usually scarce and biased.
comment: Accepted by the 28th biennial international conference on Information Processing in Medical Imaging (IPMI 2023)
☆ Audio Representation Learning by Distilling Video as Privileged Information
Deep audio representation learning using multi-modal audio-visual data often leads to a better performance compared to uni-modal approaches. However, in real-world scenarios both modalities are not always available at the time of inference, leading to performance degradation by models trained for multi-modal inference. In this work, we propose a novel approach for deep audio representation learning using audio-visual data when the video modality is absent at inference. For this purpose, we adopt teacher-student knowledge distillation under the framework of learning using privileged information (LUPI). While the previous methods proposed for LUPI use soft-labels generated by the teacher, in our proposed method we use embeddings learned by the teacher to train the student network. We integrate our method in two different settings: sequential data where the features are divided into multiple segments throughout time, and non-sequential data where the entire features are treated as one whole segment. In the non-sequential setting both the teacher and student networks are comprised of an encoder component and a task header. We use the embeddings produced by the encoder component of the teacher to train the encoder of the student, while the task header of the student is trained using ground-truth labels. In the sequential setting, the networks have an additional aggregation component that is placed between the encoder and task header. We use two sets of embeddings produced by the encoder and aggregation component of the teacher to train the student. Similar to the non-sequential setting, the task header of the student network is trained using ground-truth labels. We test our framework on two different audio-visual tasks, namely speaker recognition and speech emotion recognition and show considerable improvements over sole audio-based recognition as well as prior works that use LUPI.
☆ Uncertainty estimation for time series forecasting via Gaussian process regression surrogates
Machine learning models are widely used to solve real-world problems in science and industry. To build robust models, we should quantify the uncertainty of the model's predictions on new data. This study proposes a new method for uncertainty estimation based on the surrogate Gaussian process model. Our method can equip any base model with an accurate uncertainty estimate produced by a separate surrogate. Compared to other approaches, the estimate remains computationally effective with training only one additional model and doesn't rely on data-specific assumptions. The only requirement is the availability of the base model as a black box, which is typical. Experiments for challenging time-series forecasting data show that surrogate model-based methods provide more accurate confidence intervals than bootstrap-based methods in both medium and small-data regimes and different families of base models, including linear regression, ARIMA, and gradient boosting.
☆ Collective Robustness Certificates: Exploiting Interdependence in Graph Neural Networks ICLR 2021
In tasks like node classification, image segmentation, and named-entity recognition we have a classifier that simultaneously outputs multiple predictions (a vector of labels) based on a single input, i.e. a single graph, image, or document respectively. Existing adversarial robustness certificates consider each prediction independently and are thus overly pessimistic for such tasks. They implicitly assume that an adversary can use different perturbed inputs to attack different predictions, ignoring the fact that we have a single shared input. We propose the first collective robustness certificate which computes the number of predictions that are simultaneously guaranteed to remain stable under perturbation, i.e. cannot be attacked. We focus on Graph Neural Networks and leverage their locality property - perturbations only affect the predictions in a close neighborhood - to fuse multiple single-node certificates into a drastically stronger collective certificate. For example, on the Citeseer dataset our collective certificate for node classification increases the average number of certifiable feature perturbations from $7$ to $351$.
comment: Accepted at ICLR 2021 (https://openreview.net/forum?id=ULQdiUTHe3y). Uploaded to arxiv to fix Google Scholar indexing
☆ Migration Reframed? A multilingual analysis on the stance shift in Europe during the Ukrainian crisis
The war in Ukraine seems to have positively changed the attitude toward the critical societal topic of migration in Europe -- at least towards refugees from Ukraine. We investigate whether this impression is substantiated by how the topic is reflected in online news and social media, thus linking the representation of the issue on the Web to its perception in society. For this purpose, we combine and adapt leading-edge automatic text processing for a novel multilingual stance detection approach. Starting from 5.5M Twitter posts published by 565 European news outlets in one year, beginning September 2021, plus replies, we perform a multilingual analysis of migration-related media coverage and associated social media interaction for Europe and selected European countries. The results of our analysis show that there is actually a reframing of the discussion illustrated by the terminology change, e.g., from "migrant" to "refugee", often even accentuated with phrases such as "real refugees". However, concerning a stance shift in public perception, the picture is more diverse than expected. All analyzed cases show a noticeable temporal stance shift around the start of the war in Ukraine. Still, there are apparent national differences in the size and stability of this shift.
comment: To be published in The Web Conference 2023
☆ Federated Survival Forests
Survival analysis is a subfield of statistics concerned with modeling the occurrence time of a particular event of interest for a population. Survival analysis found widespread applications in healthcare, engineering, and social sciences. However, real-world applications involve survival datasets that are distributed, incomplete, censored, and confidential. In this context, federated learning can tremendously improve the performance of survival analysis applications. Federated learning provides a set of privacy-preserving techniques to jointly train machine learning models on multiple datasets without compromising user privacy, leading to a better generalization performance. Despite the widespread development of federated learning in recent AI research, only a few studies focus on federated survival analysis. In this work, we present a novel federated algorithm for survival analysis based on one of the most successful survival models, the random survival forest. We call the proposed method Federated Survival Forest (FedSurF). With a single communication round, FedSurF obtains a discriminative power comparable to deep-learning-based federated models trained over hundreds of federated iterations. Moreover, FedSurF retains all the advantages of random forests, namely low computational cost and natural handling of missing values and incomplete datasets. These advantages are especially desirable in real-world federated environments with multiple small datasets stored on devices with low computational capabilities. Numerical experiments compare FedSurF with state-of-the-art survival models in federated networks, showing how FedSurF outperforms deep-learning-based federated algorithms in realistic environments with non-identically distributed data.
☆ Stop overkilling simple tasks with black-box models and use transparent models instead
In recent years, the employment of deep learning methods has led to several significant breakthroughs in artificial intelligence. Different from traditional machine learning models, deep learning-based approaches are able to extract features autonomously from raw data. This allows for bypassing the feature engineering process, which is generally considered to be both error-prone and tedious. Moreover, deep learning strategies often outperform traditional models in terms of accuracy.
☆ Dealing With Non-stationarity in Decentralized Cooperative Multi-Agent Deep Reinforcement Learning via Multi-Timescale Learning
Decentralized cooperative multi-agent deep reinforcement learning (MARL) can be a versatile learning framework, particularly in scenarios where centralized training is either not possible or not practical. One of the key challenges in decentralized deep MARL is the non-stationarity of the learning environment when multiple agents are learning concurrently. A commonly used and efficient scheme for decentralized MARL is independent learning in which agents concurrently update their policies independent of each other. We first show that independent learning does not always converge, while sequential learning where agents update their policies one after another in a sequence is guaranteed to converge to an agent-by-agent optimal solution. In sequential learning, when one agent updates its policy, all other agent's policies are kept fixed, alleviating the challenge of non-stationarity due to concurrent updates in other agents' policies. However, it can be slow because only one agent is learning at any time. Therefore it might also not always be practical. In this work, we propose a decentralized cooperative MARL algorithm based on multi-timescale learning. In multi-timescale learning, all agents learn concurrently, but at different learning rates. In our proposed method, when one agent updates its policy, other agents are allowed to update their policies as well, but at a slower rate. This speeds up sequential learning, while also minimizing non-stationarity caused by other agents updating concurrently. Multi-timescale learning outperforms state-of-the-art decentralized learning methods on a set of challenging multi-agent cooperative tasks in the epymarl (papoudakis2020) benchmark. This can be seen as a first step towards more general decentralized cooperative deep MARL methods based on multi-timescale learning.
☆ A Strong Baseline for Batch Imitation Learning
Imitation of expert behaviour is a highly desirable and safe approach to the problem of sequential decision making. We provide an easy-to-implement, novel algorithm for imitation learning under a strict data paradigm, in which the agent must learn solely from data collected a priori. This paradigm allows our algorithm to be used for environments in which safety or cost are of critical concern. Our algorithm requires no additional hyper-parameter tuning beyond any standard batch reinforcement learning (RL) algorithm, making it an ideal baseline for such data-strict regimes. Furthermore, we provide formal sample complexity guarantees for the algorithm in finite Markov Decision Problems. In doing so, we formally demonstrate an unproven claim from Kearns & Singh (1998). On the empirical side, our contribution is twofold. First, we develop a practical, robust and principled evaluation protocol for offline RL methods, making use of only the dataset provided for model selection. This stands in contrast to the vast majority of previous works in offline RL, which tune hyperparameters on the evaluation environment, limiting the practical applicability when deployed in new, cost-critical environments. As such, we establish precedent for the development and fair evaluation of offline RL algorithms. Second, we evaluate our own algorithm on challenging continuous control benchmarks, demonstrating its practical applicability and competitiveness with state-of-the-art performance, despite being a simpler algorithm.
comment: 28 pages (10 main, 18 appendix), 4 figures
☆ Generative models for two-ground-truth partitions in networks
A myriad of approaches have been proposed to characterise the mesoscale structure of networks - most often as a partition based on patterns variously called communities, blocks, or clusters. Clearly, distinct methods designed to detect different types of patterns may provide a variety of answers to the network's mesoscale structure. Yet, even multiple runs of a given method can sometimes yield diverse and conflicting results, yielding entire landscapes of partitions which potentially include multiple (locally optimal) mesoscale explanations of the network. Such ambiguity motivates a closer look at the ability of these methods to find multiple qualitatively different 'ground truth' partitions in a network. Here, we propose a generative model which allows for two distinct partitions to be built into the mesoscale structure of a single benchmark network. We demonstrate a use case of the benchmark model by exploring the power of stochastic block models (SBMs) to detect coexisting bi-community and core-periphery structures of different strengths. We find that the ability to detect the two partitions individually varies considerably by SBM variant and that coexistence of both partitions is recovered only in a very limited number of cases. Our findings suggest that in most instances only one - in some way dominating - structure can be detected, even in the presence of other partitions in the generated network. They underline the need for considering entire landscapes of partitions when different competing explanations exist and motivate future research to advance partition coexistence detection methods. Our model also contributes to the field of benchmark networks more generally by enabling further exploration of the ability of new and existing methods to detect ambiguity in mesoscale structure of networks.
☆ Coherence and Diversity through Noise: Self-Supervised Paraphrase Generation via Structure-Aware Denoising
In this paper, we propose SCANING, an unsupervised framework for paraphrasing via controlled noise injection. We focus on the novel task of paraphrasing algebraic word problems having practical applications in online pedagogy as a means to reduce plagiarism as well as ensure understanding on the part of the student instead of rote memorization. This task is more complex than paraphrasing general-domain corpora due to the difficulty in preserving critical information for solution consistency of the paraphrased word problem, managing the increased length of the text and ensuring diversity in the generated paraphrase. Existing approaches fail to demonstrate adequate performance on at least one, if not all, of these facets, necessitating the need for a more comprehensive solution. To this end, we model the noising search space as a composition of contextual and syntactic aspects and sample noising functions consisting of either one or both aspects. This allows for learning a denoising function that operates over both aspects and produces semantically equivalent and syntactically diverse outputs through grounded noise injection. The denoising function serves as a foundation for learning a paraphrasing function which operates solely in the input-paraphrase space without carrying any direct dependency on noise. We demonstrate SCANING considerably improves performance in terms of both semantic preservation and producing diverse paraphrases through extensive automated and manual evaluation across 4 datasets.
comment: 12 pages (main}; 22 pages in total
The SSL Interplay: Augmentations, Inductive Bias, and Generalization
Self-supervised learning (SSL) has emerged as a powerful framework to learn representations from raw data without supervision. Yet in practice, engineers face issues such as instability in tuning optimizers and collapse of representations during training. Such challenges motivate the need for a theory to shed light on the complex interplay between the choice of data augmentation, network architecture, and training algorithm. We study such an interplay with a precise analysis of generalization performance on both pretraining and downstream tasks in a theory friendly setup, and highlight several insights for SSL practitioners that arise from our theory.
☆ Generalization Bounds with Data-dependent Fractal Dimensions
Providing generalization guarantees for modern neural networks has been a crucial task in statistical learning. Recently, several studies have attempted to analyze the generalization error in such settings by using tools from fractal geometry. While these works have successfully introduced new mathematical tools to apprehend generalization, they heavily rely on a Lipschitz continuity assumption, which in general does not hold for neural networks and might make the bounds vacuous. In this work, we address this issue and prove fractal geometry-based generalization bounds without requiring any Lipschitz assumption. To achieve this goal, we build up on a classical covering argument in learning theory and introduce a data-dependent fractal dimension. Despite introducing a significant amount of technical complications, this new notion lets us control the generalization error (over either fixed or random hypothesis spaces) along with certain mutual information (MI) terms. To provide a clearer interpretation to the newly introduced MI terms, as a next step, we introduce a notion of "geometric stability" and link our bounds to the prior art. Finally, we make a rigorous connection between the proposed data-dependent dimension and topological data analysis tools, which then enables us to compute the dimension in a numerically efficient way. We support our theory with experiments conducted on various settings.
☆ Fine-Grained Action Detection with RGB and Pose Information using Two Stream Convolutional Networks
As participants of the MediaEval 2022 Sport Task, we propose a two-stream network approach for the classification and detection of table tennis strokes. Each stream is a succession of 3D Convolutional Neural Network (CNN) blocks using attention mechanisms. Each stream processes different 4D inputs. Our method utilizes raw RGB data and pose information computed from MMPose toolbox. The pose information is treated as an image by applying the pose either on a black background or on the original RGB frame it has been computed from. Best performance is obtained by feeding raw RGB data to one stream, Pose + RGB (PRGB) information to the other stream and applying late fusion on the features. The approaches were evaluated on the provided TTStroke-21 data sets. We can report an improvement in stroke classification, reaching 87.3% of accuracy, while the detection does not outperform the baseline but still reaches an IoU of 0.349 and mAP of 0.110.
comment: Working note paper of the sport task of MediaEval 2022 in Bergen, Norway, 12-13 Jan 2023
☆ Baseline Method for the Sport Task of MediaEval 2022 with 3D CNNs using Attention Mechanisms
This paper presents the baseline method proposed for the Sports Video task part of the MediaEval 2022 benchmark. This task proposes two subtasks: stroke classification from trimmed videos, and stroke detection from untrimmed videos. This baseline addresses both subtasks. We propose two types of 3D-CNN architectures to solve the two subtasks. Both 3D-CNNs use Spatio-temporal convolutions and attention mechanisms. The architectures and the training process are tailored to solve the addressed subtask. This baseline method is shared publicly online to help the participants in their investigation and alleviate eventually some aspects of the task such as video processing, training method, evaluation and submission routine. The baseline method reaches 86.4% of accuracy with our v2 model for the classification subtask. For the detection subtask, the baseline reaches a mAP of 0.131 and IoU of 0.515 with our v1 model.
comment: Baseline paper for the sport Task of MediaEval 2022
☆ INCREASE: Inductive Graph Representation Learning for Spatio-Temporal Kriging WWW 2023
Spatio-temporal kriging is an important problem in web and social applications, such as Web or Internet of Things, where things (e.g., sensors) connected into a web often come with spatial and temporal properties. It aims to infer knowledge for (the things at) unobserved locations using the data from (the things at) observed locations during a given time period of interest. This problem essentially requires \emph{inductive learning}. Once trained, the model should be able to perform kriging for different locations including newly given ones, without retraining. However, it is challenging to perform accurate kriging results because of the heterogeneous spatial relations and diverse temporal patterns. In this paper, we propose a novel inductive graph representation learning model for spatio-temporal kriging. We first encode heterogeneous spatial relations between the unobserved and observed locations by their spatial proximity, functional similarity, and transition probability. Based on each relation, we accurately aggregate the information of most correlated observed locations to produce inductive representations for the unobserved locations, by jointly modeling their similarities and differences. Then, we design relation-aware gated recurrent unit (GRU) networks to adaptively capture the temporal correlations in the generated sequence representations for each relation. Finally, we propose a multi-relation attention mechanism to dynamically fuse the complex spatio-temporal information at different time steps from multiple relations to compute the kriging output. Experimental results on three real-world datasets show that our proposed model outperforms state-of-the-art methods consistently, and the advantage is more significant when there are fewer observed locations. Our code is available at https://github.com/zhengchuanpan/INCREASE.
comment: WWW 2023 paper
☆ Root Laplacian Eigenmaps with their application in spectral embedding
The root laplacian operator or the square root of Laplacian which can be obtained in complete Riemannian manifolds in the Gromov sense has an analog in graph theory as a square root of graph-Laplacian. Some potential applications have been shown in geometric deep learning (spectral clustering) and graph signal processing.
comment: 21 pages,4 figures
☆ Multipath agents for modular multitask ML systems
A standard ML model is commonly generated by a single method that specifies aspects such as architecture, initialization, training data and hyperparameters configuration. The presented work introduces a novel methodology allowing to define multiple methods as distinct agents. Agents can collaborate and compete to generate and improve ML models for a given tasks. The proposed methodology is demonstrated with the generation and extension of a dynamic modular multitask ML system solving more than one hundred image classification tasks. Diverse agents can compete to produce the best performing model for a task by reusing the modules introduced to the system by competing agents. The presented work focuses on the study of agents capable of: 1) reusing the modules generated by concurrent agents, 2) activating in parallel multiple modules in a frozen state by connecting them with trainable modules, 3) condition the activation mixture on each data sample by using a trainable router module. We demonstrate that this simple per-sample parallel routing method can boost the quality of the combined solutions by training a fraction of the activated parameters.
☆ Differentiable Programming of Chemical Reaction Networks
We present a differentiable formulation of abstract chemical reaction networks (CRNs) that can be trained to solve a variety of computational tasks. Chemical reaction networks are one of the most fundamental computational substrates used by nature. We study well-mixed single-chamber systems, as well as systems with multiple chambers separated by membranes, under mass-action kinetics. We demonstrate that differentiable optimisation, combined with proper regularisation, can discover non-trivial sparse reaction networks that can implement various sorts of oscillators and other chemical computing devices.
☆ Flat Seeking Bayesian Neural Networks
Bayesian Neural Networks (BNNs) offer a probabilistic interpretation for deep learning models by imposing a prior distribution over model parameters and inferencing a posterior distribution based on observed data. The model sampled from the posterior distribution can be used for providing ensemble predictions and quantifying prediction uncertainty. It is well-known that deep learning models with a lower sharpness have a better generalization ability. Nonetheless, existing posterior inferences are not aware of sharpness/flatness, hence possibly leading to high sharpness for the models sampled from it. In this paper, we develop theories, the Bayesian setting, and the variational inference approach for the sharpness-aware posterior. Specifically, the models sampled from our sharpness-aware posterior and the optimal approximate posterior estimating this sharpness-aware posterior have a better flatness, hence possibly possessing a higher generalization ability. We conduct experiments by leveraging the sharpness-aware posterior with the state-of-the-art Bayesian Neural Networks, showing that the flat-seeking counterparts outperform their baselines in all metrics of interest.
comment: Under review
☆ Network-Aided Intelligent Traffic Steering in 6G ORAN: A Multi-Layer Optimization Framework
To enable an intelligent, programmable and multi-vendor radio access network (RAN) for 6G networks, considerable efforts have been made in standardization and development of open RAN (ORAN). So far, however, the applicability of ORAN in controlling and optimizing RAN functions has not been widely investigated. In this paper, we jointly optimize the flow-split distribution, congestion control and scheduling (JFCS) to enable an intelligent traffic steering application in ORAN. Combining tools from network utility maximization and stochastic optimization, we introduce a multi-layer optimization framework that provides fast convergence, long-term utility-optimality and significant delay reduction compared to the state-of-the-art and baseline RAN approaches. Our main contributions are three-fold: i) we propose the novel JFCS framework to efficiently and adaptively direct traffic to appropriate radio units; ii) we develop low-complexity algorithms based on the reinforcement learning, inner approximation and bisection search methods to effectively solve the JFCS problem in different time scales; and iii) the rigorous theoretical performance results are analyzed to show that there exists a scaling factor to improve the tradeoff between delay and utility-optimization. Collectively, the insights in this work will open the door towards fully automated networks with enhanced control and flexibility. Numerical results are provided to demonstrate the effectiveness of the proposed algorithms in terms of the convergence rate, long-term utility-optimality and delay reduction.
comment: 33 pages, 10 figures
☆ When the Ground Truth is not True: Modelling Human Biases in Temporal Annotations
In supervised learning, low quality annotations lead to poorly performing classification and detection models, while also rendering evaluation unreliable. This is particularly apparent on temporal data, where annotation quality is affected by multiple factors. For example, in the post-hoc self-reporting of daily activities, cognitive biases are one of the most common ingredients. In particular, reporting the start and duration of an activity after its finalisation may incorporate biases introduced by personal time perceptions, as well as the imprecision and lack of granularity due to time rounding. Here we propose a method to model human biases on temporal annotations and argue for the use of soft labels. Experimental results in synthetic data show that soft labels provide a better approximation of the ground truth for several metrics. We showcase the method on a real dataset of daily activities.
☆ Languages are Rewards: Hindsight Finetuning using Human Feedback
Learning from human preferences is important for language models to be helpful and useful for humans, and to align with human and social values. Existing works focus on supervised finetuning of pretrained models, based on curated model generations that are preferred by human labelers. Such works have achieved remarkable successes in understanding and following instructions (e.g., InstructGPT, ChatGPT, etc). However, to date, a key limitation of supervised finetuning is that it cannot learn from negative ratings; models are only trained on positive-rated data, which makes it data inefficient. Because collecting human feedback data is both time consuming and expensive, it is vital for the model to learn from all feedback, akin to the remarkable ability of humans to learn from diverse feedback. In this work, we propose a novel technique called Hindsight Finetuning for making language models learn from diverse human feedback. In fact, our idea is motivated by how humans learn from hindsight experience. We condition the model on a sequence of model generations paired with hindsight feedback, and finetune the model to predict the most preferred output. By doing so, models can learn to identify and correct negative attributes or errors. Applying the method to GPT-J, we observe that it significantly improves results on summarization and dialogue tasks using the same amount of human feedback.
☆ Identifiability of latent-variable and structural-equation models: from linear to nonlinear
An old problem in multivariate statistics is that linear Gaussian models are often unidentifiable, i.e. some parameters cannot be uniquely estimated. In factor analysis, an orthogonal rotation of the factors is unidentifiable, while in linear regression, the direction of effect cannot be identified. For such linear models, non-Gaussianity of the (latent) variables has been shown to provide identifiability. In the case of factor analysis, this leads to independent component analysis, while in the case of the direction of effect, non-Gaussian versions of structural equation modelling solve the problem. More recently, we have shown how even general nonparametric nonlinear versions of such models can be estimated. Non-Gaussianity is not enough in this case, but assuming we have time series, or that the distributions are suitably modulated by some observed auxiliary variables, the models are identifiable. This paper reviews the identifiability theory for the linear and nonlinear cases, considering both factor analytic models and structural equation models.
☆ Random Forests for time-fixed and time-dependent predictors: The DynForest R package
The R package DynForest implements random forests for predicting a categorical or a (multiple causes) time-to-event outcome based on time-fixed and time-dependent predictors. Through the random forests, the time-dependent predictors can be measured with error at subject-specific times, and they can be endogeneous (i.e., impacted by the outcome process). They are modeled internally using flexible linear mixed models (thanks to lcmm package) with time-associations pre-specified by the user. DynForest computes dynamic predictions that take into account all the information from time-fixed and time-dependent predictors. DynForest also provides information about the most predictive variables using variable importance and minimal depth. Variable importance can also be computed on groups of variables. To display the results, several functions are available such as summary and plot functions. This paper aims to guide the user with a step-by-step example of the different functions for fitting random forests within DynForest.
☆ A Scalable and Efficient Iterative Method for Copying Machine Learning Classifiers
Differential replication through copying refers to the process of replicating the decision behavior of a machine learning model using another model that possesses enhanced features and attributes. This process is relevant when external constraints limit the performance of an industrial predictive system. Under such circumstances, copying enables the retention of original prediction capabilities while adapting to new demands. Previous research has focused on the single-pass implementation for copying. This paper introduces a novel sequential approach that significantly reduces the amount of computational resources needed to train or maintain a copy, leading to reduced maintenance costs for companies using machine learning models in production. The effectiveness of the sequential approach is demonstrated through experiments with synthetic and real-world datasets, showing significant reductions in time and resources, while maintaining or improving accuracy.
☆ Linking data separation, visual separation, and classifier performance using pseudo-labeling by contrastive learning
Lacking supervised data is an issue while training deep neural networks (DNNs), mainly when considering medical and biological data where supervision is expensive. Recently, Embedded Pseudo-Labeling (EPL) addressed this problem by using a non-linear projection (t-SNE) from a feature space of the DNN to a 2D space, followed by semi-supervised label propagation using a connectivity-based method (OPFSemi). We argue that the performance of the final classifier depends on the data separation present in the latent space and visual separation present in the projection. We address this by first proposing to use contrastive learning to produce the latent space for EPL by two methods (SimCLR and SupCon) and by their combination, and secondly by showing, via an extensive set of experiments, the aforementioned correlations between data separation, visual separation, and classifier performance. We demonstrate our results by the classification of five real-world challenging image datasets of human intestinal parasites with only 1% supervised samples.
comment: 10 pages, 4 figures
☆ Grounding Large Language Models in Interactive Environments with Online Reinforcement Learning
Recent works successfully leveraged Large Language Models' (LLM) abilities to capture abstract knowledge about world's physics to solve decision-making problems. Yet, the alignment between LLMs' knowledge and the environment can be wrong and limit functional competence due to lack of grounding. In this paper, we study an approach to achieve this alignment through functional grounding: we consider an agent using an LLM as a policy that is progressively updated as the agent interacts with the environment, leveraging online Reinforcement Learning to improve its performance to solve goals. Using an interactive textual environment designed to study higher-level forms of functional grounding, and a set of spatial and navigation tasks, we study several scientific questions: 1) Can LLMs boost sample efficiency for online learning of various RL tasks? 2) How can it boost different forms of generalization? 3) What is the impact of online learning? We study these questions by functionally grounding several variants (size, architecture) of FLAN-T5.
☆ L'explicabilité au service de l'extraction de connaissances : application à des données médicales
The use of machine learning has increased dramatically in the last decade. The lack of transparency is now a limiting factor, which the field of explainability wants to address. Furthermore, one of the challenges of data mining is to present the statistical relationships of a dataset when they can be highly non-linear. One of the strengths of supervised learning is its ability to find complex statistical relationships that explainability allows to represent in an intelligible way. This paper shows that explanations can be used to extract knowledge from data and shows how feature selection, data subgroup analysis and selection of highly informative instances benefit from explanations. We then present a complete data processing pipeline using these methods on medical data. -- -- L'utilisation de l'apprentissage automatique a connu un bond cette derni\`ere d\'ecennie. Le manque de transparence est aujourd'hui un frein, que le domaine de l'explicabilit\'e veut r\'esoudre. Par ailleurs, un des d\'efis de l'exploration de donn\'ees est de pr\'esenter les relations statistiques d'un jeu de donn\'ees alors que celles-ci peuvent \^etre hautement non-lin\'eaires. Une des forces de l'apprentissage supervis\'e est sa capacit\'e \`a trouver des relations statistiques complexes que l'explicabilit\'e permet de repr\'esenter de mani\`ere intelligible. Ce papier montre que les explications permettent de faire de l'extraction de connaissance sur des donn\'ees et comment la s\'election de variables, l'analyse de sous-groupes de donn\'ees et la s\'election d'instances avec un fort pouvoir informatif b\'en\'eficient des explications. Nous pr\'esentons alors un pipeline complet de traitement des donn\'ees utilisant ces m\'ethodes pour l'exploration de donn\'ees m\'edicales.
comment: 6 pages, 3 figures, EXPLAIN'AI Workshop, in French
☆ Tree-Based Learning on Amperometric Time Series Data Demonstrates High Accuracy for Classification
Elucidating exocytosis processes provide insights into cellular neurotransmission mechanisms, and may have potential in neurodegenerative diseases research. Amperometry is an established electrochemical method for the detection of neurotransmitters released from and stored inside cells. An important aspect of the amperometry method is the sub-millisecond temporal resolution of the current recordings which leads to several hundreds of gigabytes of high-quality data. In this study, we present a universal method for the classification with respect to diverse amperometric datasets using data-driven approaches in computational science. We demonstrate a very high prediction accuracy (greater than or equal to 95%). This includes an end-to-end systematic machine learning workflow for amperometric time series datasets consisting of pre-processing; feature extraction; model identification; training and testing; followed by feature importance evaluation - all implemented. We tested the method on heterogeneous amperometric time series datasets generated using different experimental approaches, chemical stimulations, electrode types, and varying recording times. We identified a certain overarching set of common features across these datasets which enables accurate predictions. Further, we showed that information relevant for the classification of amperometric traces are neither in the spiky segments alone, nor can it be retrieved from just the temporal structure of spikes. In fact, the transients between spikes and the trace baselines carry essential information for a successful classification, thereby strongly demonstrating that an effective feature representation of amperometric time series requires the full time series. To our knowledge, this is one of the first studies that propose a scheme for machine learning, and in particular, supervised learning on full amperometry time series data.
comment: 56 pages, 11 figures
☆ Hybrid Contrastive Constraints for Multi-Scenario Ad Ranking
Multi-scenario ad ranking aims at leveraging the data from multiple domains or channels for training a unified ranking model to improve the performance at each individual scenario. Although the research on this task has made important progress, it still lacks the consideration of cross-scenario relations, thus leading to limitation in learning capability and difficulty in interrelation modeling. In this paper, we propose a Hybrid Contrastive Constrained approach (HC^2) for multi-scenario ad ranking. To enhance the modeling of data interrelation, we elaborately design a hybrid contrastive learning approach to capture commonalities and differences among multiple scenarios. The core of our approach consists of two elaborated contrastive losses, namely generalized and individual contrastive loss, which aim at capturing common knowledge and scenario-specific knowledge, respectively. To adapt contrastive learning to the complex multi-scenario setting, we propose a series of important improvements. For generalized contrastive loss, we enhance contrastive learning by extending the contrastive samples (label-aware and diffusion noise enhanced contrastive samples) and reweighting the contrastive samples (reciprocal similarity weighting). For individual contrastive loss, we use the strategies of dropout-based augmentation and {cross-scenario encoding} for generating meaningful positive and negative contrastive samples, respectively. Extensive experiments on both offline evaluation and online test have demonstrated the effectiveness of the proposed HC$^2$ by comparing it with a number of competitive baselines.
comment: 10 pages, 5 figures
☆ Trust, but Verify: Using Self-Supervised Probing to Improve Trustworthiness
Trustworthy machine learning is of primary importance to the practical deployment of deep learning models. While state-of-the-art models achieve astonishingly good performance in terms of accuracy, recent literature reveals that their predictive confidence scores unfortunately cannot be trusted: e.g., they are often overconfident when wrong predictions are made, or so even for obvious outliers. In this paper, we introduce a new approach of self-supervised probing, which enables us to check and mitigate the overconfidence issue for a trained model, thereby improving its trustworthiness. We provide a simple yet effective framework, which can be flexibly applied to existing trustworthiness-related methods in a plug-and-play manner. Extensive experiments on three trustworthiness-related tasks (misclassification detection, calibration and out-of-distribution detection) across various benchmarks verify the effectiveness of our proposed probing framework.
comment: European Conference on Computer Vision 2022
☆ Uncertainty Calibration and its Application to Object Detection
Image-based environment perception is an important component especially for driver assistance systems or autonomous driving. In this scope, modern neuronal networks are used to identify multiple objects as well as the according position and size information within a single frame. The performance of such an object detection model is important for the overall performance of the whole system. However, a detection model might also predict these objects under a certain degree of uncertainty. [...] In this work, we examine the semantic uncertainty (which object type?) as well as the spatial uncertainty (where is the object and how large is it?). We evaluate if the predicted uncertainties of an object detection model match with the observed error that is achieved on real-world data. In the first part of this work, we introduce the definition for confidence calibration of the semantic uncertainty in the context of object detection, instance segmentation, and semantic segmentation. We integrate additional position information in our examinations to evaluate the effect of the object's position on the semantic calibration properties. Besides measuring calibration, it is also possible to perform a post-hoc recalibration of semantic uncertainty that might have turned out to be miscalibrated. [...] The second part of this work deals with the spatial uncertainty obtained by a probabilistic detection model. [...] We review and extend common calibration methods so that it is possible to obtain parametric uncertainty distributions for the position information in a more flexible way. In the last part, we demonstrate a possible use-case for our derived calibration methods in the context of object tracking. [...] We integrate our previously proposed calibration techniques and demonstrate the usefulness of semantic and spatial uncertainty calibration in a subsequent process. [...]
comment: PhD thesis at University of Wuppertal, cite by: 'Fabian K\"uppers. "Uncertainty Calibration and its Application to Object Detection." PhD Thesis, University of Wuppertal, January 2023'
☆ COVID-19 Infection Analysis Framework using Novel Boosted CNNs and Radiological Images
COVID-19 is a new pathogen that first appeared in the human population at the end of 2019, and it can lead to novel variants of pneumonia after infection. COVID-19 is a rapidly spreading infectious disease that infects humans faster. Therefore, efficient diagnostic systems may accurately identify infected patients and thus help control their spread. In this regard, a new two-stage analysis framework is developed to analyze minute irregularities of COVID-19 infection. A novel detection Convolutional Neural Network (CNN), STM-BRNet, is developed that incorporates the Split-Transform-Merge (STM) block and channel boosting (CB) to identify COVID-19 infected CT slices in the first stage. Each STM block extracts boundary and region-smoothing-specific features for COVID-19 infection detection. Moreover, the various boosted channels are obtained by introducing the new CB and Transfer Learning (TL) concept in STM blocks to capture small illumination and texture variations of COVID-19-specific images. The COVID-19 CTs are provided with new SA-CB-BRSeg segmentation CNN for delineating infection in images in the second stage. SA-CB-BRSeg methodically utilized smoothening and heterogeneous operations in the encoder and decoder to capture simultaneously COVID-19 specific patterns that are region homogeneity, texture variation, and boundaries. Additionally, the new CB concept is introduced in the decoder of SA-CB-BRSeg by combining additional channels using TL to learn the low contrast region. The proposed STM-BRNet and SA-CB-BRSeg yield considerable achievement in accuracy: 98.01 %, Recall: 98.12%, F-score: 98.11%, and Dice Similarity: 96.396%, IOU: 98.845 % for the COVID-19 infectious region, respectively. The proposed two-stage framework significantly increased performance compared to single-phase and other reported systems and reduced the burden on the radiologists.
comment: 26 Pages, 11 Figures, 6 Tables. arXiv admin note: text overlap with arXiv:2209.10963
☆ Leveraging Domain Relations for Domain Generalization
Distribution shift is a major challenge in machine learning, as models often perform poorly during the test stage if the test distribution differs from the training distribution. In this paper, we focus on domain shifts, which occur when the model is applied to new domains that are different from the ones it was trained on, and propose a new approach called D^3G. Unlike previous approaches that aim to learn a single model that is domain invariant, D^3G learns domain-specific models by leveraging the relations among different domains. Concretely, D^3G learns a set of training-domain-specific functions during the training stage and reweights them based on domain relations during the test stage. These domain relations can be directly derived or learned from fixed domain meta-data. Under mild assumptions, we theoretically proved that using domain relations to reweight training-domain-specific functions achieves stronger generalization compared to averaging them. Empirically, we evaluated the effectiveness of D^3G using both toy and real-world datasets for tasks such as temperature regression, land use classification, and molecule-protein interaction prediction. Our results showed that D^3G consistently outperformed state-of-the-art methods, with an average improvement of 10.6% in performance.
☆ Target-based Surrogates for Stochastic Optimization
We consider minimizing functions for which it is expensive to compute the (possibly stochastic) gradient. Such functions are prevalent in reinforcement learning, imitation learning and adversarial training. Our target optimization framework uses the (expensive) gradient computation to construct surrogate functions in a target space (e.g. the logits output by a linear model for classification) that can be minimized efficiently. This allows for multiple parameter updates to the model, amortizing the cost of gradient computation. In the full-batch setting, we prove that our surrogate is a global upper-bound on the loss, and can be (locally) minimized using a black-box optimization algorithm. We prove that the resulting majorization-minimization algorithm ensures convergence to a stationary point of the loss. Next, we instantiate our framework in the stochastic setting and propose the $SSO$ algorithm, which can be viewed as projected stochastic gradient descent in the target space. This connection enables us to prove theoretical guarantees for $SSO$ when minimizing convex functions. Our framework allows the use of standard stochastic optimization algorithms to construct surrogates which can be minimized by any deterministic optimization method. To evaluate our framework, we consider a suite of supervised learning and imitation learning problems. Our experiments indicate the benefits of target optimization and the effectiveness of $SSO$.
☆ Toward Large Kernel Models
Recent studies indicate that kernel machines can often perform similarly or better than deep neural networks (DNNs) on small datasets. The interest in kernel machines has been additionally bolstered by the discovery of their equivalence to wide neural networks in certain regimes. However, a key feature of DNNs is their ability to scale the model size and training data size independently, whereas in traditional kernel machines model size is tied to data size. Because of this coupling, scaling kernel machines to large data has been computationally challenging. In this paper, we provide a way forward for constructing large-scale general kernel models, which are a generalization of kernel machines that decouples the model and data, allowing training on large datasets. Specifically, we introduce EigenPro 3.0, an algorithm based on projected dual preconditioned SGD and show scaling to model and data sizes which have not been possible with existing kernel methods.
comment: Code is available at github.com/EigenPro/EigenPro3
☆ Learning Representations of Bi-Level Knowledge Graphs for Reasoning beyond Link Prediction AAAI
Knowledge graphs represent known facts using triplets. While existing knowledge graph embedding methods only consider the connections between entities, we propose considering the relationships between triplets. For example, let us consider two triplets $T_1$ and $T_2$ where $T_1$ is (Academy_Awards, Nominates, Avatar) and $T_2$ is (Avatar, Wins, Academy_Awards). Given these two base-level triplets, we see that $T_1$ is a prerequisite for $T_2$. In this paper, we define a higher-level triplet to represent a relationship between triplets, e.g., $\langle T_1$, PrerequisiteFor, $T_2\rangle$ where PrerequisiteFor is a higher-level relation. We define a bi-level knowledge graph that consists of the base-level and the higher-level triplets. We also propose a data augmentation strategy based on the random walks on the bi-level knowledge graph to augment plausible triplets. Our model called BiVE learns embeddings by taking into account the structures of the base-level and the higher-level triplets, with additional consideration of the augmented triplets. We propose two new tasks: triplet prediction and conditional link prediction. Given a triplet $T_1$ and a higher-level relation, the triplet prediction predicts a triplet that is likely to be connected to $T_1$ by the higher-level relation, e.g., $\langle T_1$, PrerequisiteFor, ?$\rangle$. The conditional link prediction predicts a missing entity in a triplet conditioned on another triplet, e.g., $\langle T_1$, PrerequisiteFor, (Avatar, Wins, ?)$\rangle$. Experimental results show that BiVE significantly outperforms all other methods in the two new tasks and the typical base-level link prediction in real-world bi-level knowledge graphs.
comment: 14 pages, 3 figures, 15 tables. 37th AAAI Conference on Artificial Intelligence (AAAI 2023)
☆ MAP: Memory-aware Automated Intra-op Parallel Training For Foundation Models
Recently, large models have achieved the state of the art performances in various fields. In order to support large model training, we have to use distributed training techniques. However, finding an efficient distributed execution plan not only requires fine-grained model statistics, such as memory and computing overhead of each operator but also is a labor-intensive task even for an expert in the field of distributed training. In this paper, we introduce MAP, a compiler built upon PyTorch to implement Memory-aware Automated Parallelization. To profiling operator costs, existing training systems and machine learning pipelines either physically execute with respect to each operand or estimate the memory usage with a scaled input tensor, which are often time-consuming and misleading. Compared with existing methods, MAP provides an easy-to-use symbolic profiler to generate memory and computing statistics of an arbitrary PyTorch model with trivial time cost, so it will boost high productivity for ML developers. In addition, MAP can also seamlessly speed up different static planning tasks on computation graphs for PyTorch, and requires only a few lines of modification to user code to generate a new module instance that has a top-performing distributed execution plan. The source code is publicly available at https://github.com/hpcaitech/ColossalAI
☆ ProbPNN: Enhancing Deep Probabilistic Forecasting with Statistical Information
Probabilistic forecasts are essential for various downstream applications such as business development, traffic planning, and electrical grid balancing. Many of these probabilistic forecasts are performed on time series data that contain calendar-driven periodicities. However, existing probabilistic forecasting methods do not explicitly take these periodicities into account. Therefore, in the present paper, we introduce a deep learning-based method that considers these calendar-driven periodicities explicitly. The present paper, thus, has a twofold contribution: First, we apply statistical methods that use calendar-driven prior knowledge to create rolling statistics and combine them with neural networks to provide better probabilistic forecasts. Second, we benchmark ProbPNN with state-of-the-art benchmarks by comparing the achieved normalised continuous ranked probability score (nCRPS) and normalised Pinball Loss (nPL) on two data sets containing in total more than 1000 time series. The results of the benchmarks show that using statistical forecasting components improves the probabilistic forecast performance and that ProbPNN outperforms other deep learning forecasting methods whilst requiring less computation costs.
☆ Ten Lessons We Have Learned in the New "Sparseland": A Short Handbook for Sparse Neural Network Researchers
This article does not propose any novel algorithm or new hardware for sparsity. Instead, it aims to serve the "common good" for the increasingly prosperous Sparse Neural Network (SNN) research community. We attempt to summarize some most common confusions in SNNs, that one may come across in various scenarios such as paper review/rebuttal and talks - many drawn from the authors' own bittersweet experiences! We feel that doing so is meaningful and timely, since the focus of SNN research is notably shifting from traditional pruning to more diverse and profound forms of sparsity before, during, and after training. The intricate relationships between their scopes, assumptions, and approaches lead to misunderstandings, for non-experts or even experts in SNNs. In response, we summarize ten Q\&As of SNNs from many key aspects, including dense vs. sparse, unstructured sparse vs. structured sparse, pruning vs. sparse training, dense-to-sparse training vs. sparse-to-sparse training, static sparsity vs. dynamic sparsity, before-training/during-training vs. post-training sparsity, and many more. We strive to provide proper and generically applicable answers to clarify those confusions to the best extent possible. We hope our summary provides useful general knowledge for people who want to enter and engage with this exciting community; and also provides some "mind of ease" convenience for SNN researchers to explain their work in the right contexts. At the very least (and perhaps as this article's most insignificant target functionality), if you are writing/planning to write a paper or rebuttal in the field of SNNs, we hope some of our answers could help you!
☆ Clarifying Trust of Materials Property Predictions using Neural Networks with Distribution-Specific Uncertainty Quantification
It is critical that machine learning (ML) model predictions be trustworthy for high-throughput catalyst discovery approaches. Uncertainty quantification (UQ) methods allow estimation of the trustworthiness of an ML model, but these methods have not been well explored in the field of heterogeneous catalysis. Herein, we investigate different UQ methods applied to a crystal graph convolutional neural network (CGCNN) to predict adsorption energies of molecules on alloys from the Open Catalyst 2020 (OC20) dataset, the largest existing heterogeneous catalyst dataset. We apply three UQ methods to the adsorption energy predictions, namely k-fold ensembling, Monte Carlo dropout, and evidential regression. The effectiveness of each UQ method is assessed based on accuracy, sharpness, dispersion, calibration, and tightness. Evidential regression is demonstrated to be a powerful approach for rapidly obtaining tunable, competitively trustworthy UQ estimates for heterogeneous catalysis applications when using neural networks. Recalibration of model uncertainties is shown to be essential in practical screening applications of catalysts using uncertainties.
comment: 28 pages, 16 figures (8 main text, 8 SI), submitted to Machine Learning: Science & Technology journal (MLST, IOP)
☆ Generative Diffusion Models on Graphs: Methods and Applications
Diffusion models, as a novel generative paradigm, have achieved remarkable success in various image generation tasks such as image inpainting, image-to-text translation, and video generation. Graph generation is a crucial computational task on graphs with numerous real-world applications. It aims to learn the distribution of given graphs and then generate new graphs. Given the great success of diffusion models in image generation, increasing efforts have been made to leverage these techniques to advance graph generation in recent years. In this paper, we first provide a comprehensive overview of generative diffusion models on graphs, In particular, we review representative algorithms for three variants of graph diffusion models, i.e., Score Matching with Langevin Dynamics (SMLD), Denoising Diffusion Probabilistic Model (DDPM), and Score-based Generative Model (SGM). Then, we summarize the major applications of generative diffusion models on graphs with a specific focus on molecule and protein modeling. Finally, we discuss promising directions in generative diffusion models on graph-structured data.
☆ $z$-SignFedAvg: A Unified Stochastic Sign-based Compression for Federated Learning
Federated Learning (FL) is a promising privacy-preserving distributed learning paradigm but suffers from high communication cost when training large-scale machine learning models. Sign-based methods, such as SignSGD \cite{bernstein2018signsgd}, have been proposed as a biased gradient compression technique for reducing the communication cost. However, sign-based algorithms could diverge under heterogeneous data, which thus motivated the development of advanced techniques, such as the error-feedback method and stochastic sign-based compression, to fix this issue. Nevertheless, these methods still suffer from slower convergence rates. Besides, none of them allows multiple local SGD updates like FedAvg \cite{mcmahan2017communication}. In this paper, we propose a novel noisy perturbation scheme with a general symmetric noise distribution for sign-based compression, which not only allows one to flexibly control the tradeoff between gradient bias and convergence performance, but also provides a unified viewpoint to existing stochastic sign-based methods. More importantly, the unified noisy perturbation scheme enables the development of the very first sign-based FedAvg algorithm ($z$-SignFedAvg) to accelerate the convergence. Theoretically, we show that $z$-SignFedAvg achieves a faster convergence rate than existing sign-based methods and, under the uniformly distributed noise, can enjoy the same convergence rate as its uncompressed counterpart. Extensive experiments are conducted to demonstrate that the $z$-SignFedAvg can achieve competitive empirical performance on real datasets and outperforms existing schemes.
☆ Topology-aware Federated Learning in Edge Computing: A Comprehensive Survey
The ultra-low latency requirements of 5G/6G applications and privacy constraints call for distributed machine learning systems to be deployed at the edge. With its simple yet effective approach, federated learning (FL) is proved to be a natural solution for massive user-owned devices in edge computing with distributed and private training data. Most vanilla FL algorithms based on FedAvg follow a naive star topology, ignoring the heterogeneity and hierarchy of the volatile edge computing architectures and topologies in reality. In this paper, we conduct a comprehensive survey on the existing work of optimized FL models, frameworks, and algorithms with a focus on their network topologies. After a brief recap of FL and edge computing networks, we introduce various types of edge network topologies, along with the optimizations under the aforementioned network topologies. Lastly, we discuss the remaining challenges and future works for applying FL in topology-specific edge networks.
comment: 36 pages, 16 figures, journal submission
♻ ☆ A survey on knowledge-enhanced multimodal learning
Multimodal learning has been a field of increasing interest, aiming to combine various modalities in a single joint representation. Especially in the area of visiolinguistic (VL) learning multiple models and techniques have been developed, targeting a variety of tasks that involve images and text. VL models have reached unprecedented performances by extending the idea of Transformers, so that both modalities can learn from each other. Massive pre-training procedures enable VL models to acquire a certain level of real-world understanding, although many gaps can be identified: the limited comprehension of commonsense, factual, temporal and other everyday knowledge aspects questions the extendability of VL tasks. Knowledge graphs and other knowledge sources can fill those gaps by explicitly providing missing information, unlocking novel capabilities of VL models. In the same time, knowledge graphs enhance explainability, fairness and validity of decision making, issues of outermost importance for such complex implementations. The current survey aims to unify the fields of VL representation learning and knowledge graphs, and provides a taxonomy and analysis of knowledge-enhanced VL models.
♻ ☆ The mbsts package: Multivariate Bayesian Structural Time Series Models in R
The multivariate Bayesian structural time series (MBSTS) model is a general machine learning model that deals with inference and prediction for multiple correlated time series, where one also has the choice of using a different candidate pool of contemporaneous predictors for each target series. The MBSTS model has wide applications and is ideal for feature selection, time series forecasting, nowcasting, inferring causal impact, and others. This paper demonstrates how to use the R package mbsts for MBSTS modeling, establishing a bridge between user-friendly and developer-friendly functions in the package and the corresponding methodology. Object-oriented functions in the package are explained in the way that enables users to flexibly add or deduct some components, as well as to simplify or complicate some settings.
♻ ☆ Fast, Differentiable and Sparse Top-k: a Convex Analysis Perspective
The top-k operator returns a k-sparse vector, where the non-zero values correspond to the k largest values of the input. Unfortunately, because it is a discontinuous function, it is difficult to incorporate in neural networks trained end-to-end with backpropagation. Recent works have considered differentiable relaxations, based either on regularization or perturbation techniques. However, to date, no approach is fully differentiable and sparse. In this paper, we propose new differentiable and sparse top-k operators. We view the top-k operator as a linear program over the permutahedron, the convex hull of permutations. We then introduce a p-norm regularization term to smooth out the operator, and show that its computation can be reduced to isotonic optimization. Our framework is significantly more general than the existing one and allows for example to express top-k operators that select values in magnitude. On the algorithmic side, in addition to pool adjacent violator (PAV) algorithms, we propose a new GPU/TPU-friendly Dykstra algorithm to solve isotonic optimization problems. We successfully use our operators to prune weights in neural networks, to fine-tune vision transformers, and as a router in sparse mixture of experts.
comment: 23 pages
♻ ☆ Improving Few-Shot Generalization by Exploring and Exploiting Auxiliary Data
Few-shot learning involves learning an effective model from only a few labeled datapoints. The use of a small training set makes it difficult to avoid overfitting but also makes few-shot learning applicable to many important real-world settings. In this work, we focus on Few-shot Learning with Auxiliary Data (FLAD), a training paradigm that assumes access to auxiliary data during few-shot learning in hopes of improving generalization. Introducing auxiliary data during few-shot learning leads to essential design choices where hand-designed heuristics can lead to sub-optimal performance. In this work, we focus on automated sampling strategies for FLAD and relate them to the explore-exploit dilemma that is central in multi-armed bandit settings. Based on this connection we propose two algorithms -- EXP3-FLAD and UCB1-FLAD -- and compare them with methods that either explore or exploit, finding that the combination of exploration and exploitation is crucial. Using our proposed algorithms to train T5 yields a 9% absolute improvement over the explicitly multi-task pre-trained T0 model across 11 datasets.
comment: 19 pages, 7 figures, code available at https://github.com/alon-albalak/FLAD
♻ ☆ ByteTransformer: A High-Performance Transformer Boosted for Variable-Length Inputs
Transformers have become keystone models in natural language processing over the past decade. They have achieved great popularity in deep learning applications, but the increasing sizes of the parameter spaces required by transformer models generate a commensurate need to accelerate performance. Natural language processing problems are also routinely faced with variable-length sequences, as word counts commonly vary among sentences. Existing deep learning frameworks pad variable-length sequences to a maximal length, which adds significant memory and computational overhead. In this paper, we present ByteTransformer, a high-performance transformer boosted for variable-length inputs. We propose a padding-free algorithm that liberates the entire transformer from redundant computations on zero padded tokens. In addition to algorithmic-level optimization, we provide architecture-aware optimizations for transformer functional modules, especially the performance-critical algorithm Multi-Head Attention (MHA). Experimental results on an NVIDIA A100 GPU with variable-length sequence inputs validate that our fused MHA outperforms PyTorch by 6.13x. The end-to-end performance of ByteTransformer for a forward BERT transformer surpasses state-of-the-art transformer frameworks, such as PyTorch JIT, TensorFlow XLA, Tencent TurboTransformer, Microsoft DeepSpeed-Inference and NVIDIA FasterTransformer, by 87\%, 131\%, 138\%, 74\% and 55\%, respectively. We also demonstrate the general applicability of our optimization methods to other BERT-like models, including ALBERT, DistilBERT, and DeBERTa.
comment: Accepted at IPDPS 2023
♻ ☆ Deep Dependency Networks for Multi-Label Classification
We propose a simple approach which combines the strengths of probabilistic graphical models and deep learning architectures for solving the multi-label classification task, focusing specifically on image and video data. First, we show that the performance of previous approaches that combine Markov Random Fields with neural networks can be modestly improved by leveraging more powerful methods such as iterative join graph propagation, integer linear programming, and $\ell_1$ regularization-based structure learning. Then we propose a new modeling framework called deep dependency networks, which augments a dependency network, a model that is easy to train and learns more accurate dependencies but is limited to Gibbs sampling for inference, to the output layer of a neural network. We show that despite its simplicity, jointly learning this new architecture yields significant improvements in performance over the baseline neural network. In particular, our experimental evaluation on three video activity classification datasets: Charades, Textually Annotated Cooking Scenes (TACoS), and Wetlab, and three multi-label image classification datasets: MS-COCO, PASCAL VOC, and NUS-WIDE show that deep dependency networks are almost always superior to pure neural architectures that do not use dependency networks.
♻ ☆ Switchable Lightweight Anti-symmetric Processing (SLAP) with CNN Outspeeds Data Augmentation by Smaller Sample -- Application in Gomoku Reinforcement Learning
To replace data augmentation, this paper proposed a method called SLAP to intensify experience to speed up machine learning and reduce the sample size. SLAP is a model-independent protocol/function to produce the same output given different transformation variants. SLAP improved the convergence speed of convolutional neural network learning by 83% in the experiments with Gomoku game states, with only one eighth of the sample size compared with data augmentation. In reinforcement learning for Gomoku, using AlphaGo Zero/AlphaZero algorithm with data augmentation as baseline, SLAP reduced the number of training samples by a factor of 8 and achieved similar winning rate against the same evaluator, but it was not yet evident that it could speed up reinforcement learning. The benefits should at least apply to domains that are invariant to symmetry or certain transformations. As future work, SLAP may aid more explainable learning and transfer learning for domains that are not invariant to symmetry, as a small step towards artificial general intelligence.
comment: Add co-author and enrich discussion; 6 pages, 8 figures
♻ ☆ Confidence-Ranked Reconstruction of Census Microdata from Published Statistics
A reconstruction attack on a private dataset $D$ takes as input some publicly accessible information about the dataset and produces a list of candidate elements of $D$. We introduce a new class of data reconstruction attacks based on randomized methods for non-convex optimization. We empirically demonstrate that our attacks can not only reconstruct full rows of $D$ from aggregate query statistics $Q(D)\in \mathbb{R}^m$, but can do so in a way that reliably ranks reconstructed rows by their odds of appearing in the private data, providing a signature that could be used for prioritizing reconstructed rows for further actions such as identify theft or hate crime. We also design a sequence of baselines for evaluating reconstruction attacks. Our attacks significantly outperform those that are based only on access to a public distribution or population from which the private dataset $D$ was sampled, demonstrating that they are exploiting information in the aggregate statistics $Q(D)$, and not simply the overall structure of the distribution. In other words, the queries $Q(D)$ are permitting reconstruction of elements of this dataset, not the distribution from which $D$ was drawn. These findings are established both on 2010 U.S. decennial Census data and queries and Census-derived American Community Survey datasets. Taken together, our methods and experiments illustrate the risks in releasing numerically precise aggregate statistics of a large dataset, and provide further motivation for the careful application of provably private techniques such as differential privacy.
♻ ☆ The KFIoU Loss for Rotated Object Detection ICLR 2023
Differing from the well-developed horizontal object detection area whereby the computing-friendly IoU based loss is readily adopted and well fits with the detection metrics. In contrast, rotation detectors often involve a more complicated loss based on SkewIoU which is unfriendly to gradient-based training. In this paper, we propose an effective approximate SkewIoU loss based on Gaussian modeling and Gaussian product, which mainly consists of two items. The first term is a scale-insensitive center point loss, which is used to quickly narrow the distance between the center points of the two bounding boxes. In the distance-independent second term, the product of the Gaussian distributions is adopted to inherently mimic the mechanism of SkewIoU by its definition, and show its alignment with the SkewIoU loss at trend-level within a certain distance (i.e. within 9 pixels). This is in contrast to recent Gaussian modeling based rotation detectors e.g. GWD loss and KLD loss that involve a human-specified distribution distance metric which require additional hyperparameter tuning that vary across datasets and detectors. The resulting new loss called KFIoU loss is easier to implement and works better compared with exact SkewIoU loss, thanks to its full differentiability and ability to handle the non-overlapping cases. We further extend our technique to the 3-D case which also suffers from the same issues as 2-D. Extensive results on various public datasets (2-D/3-D, aerial/text/face images) with different base detectors show the effectiveness of our approach.
comment: 18 pages, 6 figures, 8 tables, accepted by ICLR 2023, TensorFlow code: https://github.com/yangxue0827/RotationDetection, PyTorch code: https://github.com/open-mmlab/mmrotate, Jittor code: https://github.com/Jittor/JDet
♻ ☆ Unsupervised hierarchical clustering using the learning dynamics of RBMs
Datasets in the real world are often complex and to some degree hierarchical, with groups and sub-groups of data sharing common characteristics at different levels of abstraction. Understanding and uncovering the hidden structure of these datasets is an important task that has many practical applications. To address this challenge, we present a new and general method for building relational data trees by exploiting the learning dynamics of the Restricted Boltzmann Machine (RBM). Our method is based on the mean-field approach, derived from the Plefka expansion, and developed in the context of disordered systems. It is designed to be easily interpretable. We tested our method in an artificially created hierarchical dataset and on three different real-world datasets (images of digits, mutations in the human genome, and a homologous family of proteins). The method is able to automatically identify the hierarchical structure of the data. This could be useful in the study of homologous protein sequences, where the relationships between proteins are critical for understanding their function and evolution.
comment: 33 pages, 17 figures
♻ ☆ Learning Variational Models with Unrolling and Bilevel Optimization
In this paper we consider the problem learning of variational models in the context of supervised learning via risk minimization. Our goal is to provide a deeper understanding of the two approaches of learning of variational models via bilevel optimization and via algorithm unrolling. The former considers the variational model as a lower level optimization problem below the risk minimization problem, while the latter replaces the lower level optimization problem by an algorithm that solves said problem approximately. Both approaches are used in practice, but, unrolling is much simpler from a computational point of view. To analyze and compare the two approaches, we consider a simple toy model, and compute all risks and the respective estimators explicitly. We show that unrolling can be better than the bilevel optimization approach, but also that the performance of unrolling can depend significantly on further parameters, sometimes in unexpected ways: While the stepsize of the unrolled algorithm matters a lot, the number of unrolled iterations only matters if the number is even or odd, and these two cases are notably different.
♻ ☆ Learning Mixtures of Markov Chains and MDPs
We present an algorithm for learning mixtures of Markov chains and Markov decision processes (MDPs) from short unlabeled trajectories. Specifically, our method handles mixtures of Markov chains with optional control input by going through a multi-step process, involving (1) a subspace estimation step, (2) spectral clustering of trajectories using "pairwise distance estimators," along with refinement using the EM algorithm, (3) a model estimation step, and (4) a classification step for predicting labels of new trajectories. We provide end-to-end performance guarantees, where we only explicitly require the length of trajectories to be linear in the number of states and the number of trajectories to be linear in a mixing time parameter. Experimental results support these guarantees, where we attain 96.6% average accuracy on a mixture of two MDPs in gridworld, outperforming the EM algorithm with random initialization (73.2% average accuracy).
comment: 51 pages (13 page paper, 38 page appendix). Paper restructured and refined, corrections made to proofs, experiments added
♻ ☆ On the complexity of nonsmooth automatic differentiation
Using the notion of conservative gradient, we provide a simple model to estimate the computational costs of the backward and forward modes of algorithmic differentiation for a wide class of nonsmooth programs. The overhead complexity of the backward mode turns out to be independent of the dimension when using programs with locally Lipschitz semi-algebraic or definable elementary functions. This considerably extends Baur-Strassen's smooth cheap gradient principle. We illustrate our results by establishing fast backpropagation results of conservative gradients through feedforward neural networks with standard activation and loss functions. Nonsmooth backpropagation's cheapness contrasts with concurrent forward approaches, which have, to this day, dimensional-dependent worst-case overhead estimates. We provide further results suggesting the superiority of backward propagation of conservative gradients. Indeed, we relate the complexity of computing a large number of directional derivatives to that of matrix multiplication, and we show that finding two subgradients in the Clarke subdifferential of a function is an NP-hard problem.
♻ ☆ Iterative Deepening Hyperband
Hyperparameter optimization (HPO) is concerned with the automated search for the most appropriate hyperparameter configuration (HPC) of a parameterized machine learning algorithm. A state-of-the-art HPO method is Hyperband, which, however, has its own parameters that influence its performance. One of these parameters, the maximal budget, is especially problematic: If chosen too small, the budget needs to be increased in hindsight and, as Hyperband is not incremental by design, the entire algorithm must be re-run. This is not only costly but also comes with a loss of valuable knowledge already accumulated. In this paper, we propose incremental variants of Hyperband that eliminate these drawbacks, and show that these variants satisfy theoretical guarantees qualitatively similar to those for the original Hyperband with the "right" budget. Moreover, we demonstrate their practical utility in experiments with benchmark data sets.
♻ ☆ Censored Quantile Regression Neural Networks for Distribution-Free Survival Analysis NeurIPS 2022
This paper considers doing quantile regression on censored data using neural networks (NNs). This adds to the survival analysis toolkit by allowing direct prediction of the target variable, along with a distribution-free characterisation of uncertainty, using a flexible function approximator. We begin by showing how an algorithm popular in linear models can be applied to NNs. However, the resulting procedure is inefficient, requiring sequential optimisation of an individual NN at each desired quantile. Our major contribution is a novel algorithm that simultaneously optimises a grid of quantiles output by a single NN. To offer theoretical insight into our algorithm, we show firstly that it can be interpreted as a form of expectation-maximisation, and secondly that it exhibits a desirable `self-correcting' property. Experimentally, the algorithm produces quantiles that are better calibrated than existing methods on 10 out of 12 real datasets.
comment: Published in NeurIPS 2022
♻ ☆ Policy-Value Alignment and Robustness in Search-based Multi-Agent Learning
Large-scale AI systems that combine search and learning have reached super-human levels of performance in game-playing, but have also been shown to fail in surprising ways. The brittleness of such models limits their efficacy and trustworthiness in real-world deployments. In this work, we systematically study one such algorithm, AlphaZero, and identify two phenomena related to the nature of exploration. First, we find evidence of policy-value misalignment -- for many states, AlphaZero's policy and value predictions contradict each other, revealing a tension between accurate move-selection and value estimation in AlphaZero's objective. Further, we find inconsistency within AlphaZero's value function, which causes it to generalize poorly, despite its policy playing an optimal strategy. From these insights we derive VISA-VIS: a novel method that improves policy-value alignment and value robustness in AlphaZero. Experimentally, we show that our method reduces policy-value misalignment by up to 76%, reduces value generalization error by up to 50%, and reduces average value error by up to 55%.
comment: 9 pages, 5 figures
♻ ☆ Reasoning Through Memorization: Nearest Neighbor Knowledge Graph Embeddings
Previous knowledge graph embedding approaches usually map entities to representations and utilize score functions to predict the target entities, yet they typically struggle to reason rare or emerging unseen entities. In this paper, we propose kNN-KGE, a new knowledge graph embedding approach with pre-trained language models, by linearly interpolating its entity distribution with k-nearest neighbors. We compute the nearest neighbors based on the distance in the entity embedding space from the knowledge store. Our approach can allow rare or emerging entities to be memorized explicitly rather than implicitly in model parameters. Experimental results demonstrate that our approach can improve inductive and transductive link prediction results and yield better performance for low-resource settings with only a few triples, which might be easier to reason via explicit memory. Code is available at https://github.com/zjunlp/KNN-KG.
comment: Work in progress
♻ ☆ FewShotTextGCN: K-hop neighborhood regularization for few-shot learning on graphs EACL 2023
We present FewShotTextGCN, a novel method designed to effectively utilize the properties of word-document graphs for improved learning in low-resource settings. We introduce K-hop Neighbourhood Regularization, a regularizer for heterogeneous graphs, and show that it stabilizes and improves learning when only a few training samples are available. We furthermore propose a simplification in the graph-construction method, which results in a graph that is $\sim$7 times less dense and yields better performance in little-resource settings while performing on par with the state of the art in high-resource settings. Finally, we introduce a new variant of Adaptive Pseudo-Labeling tailored for word-document graphs. When using as little as 20 samples for training, we outperform a strong TextGCN baseline with 17% in absolute accuracy on average over eight languages. We demonstrate that our method can be applied to document classification without any language model pretraining on a wide range of typologically diverse languages while performing on par with large pretrained language models.
comment: 8 pages, 4 figures, EACL 2023
♻ ☆ Hierarchical Sliced Wasserstein Distance ICLR 2023
Sliced Wasserstein (SW) distance has been widely used in different application scenarios since it can be scaled to a large number of supports without suffering from the curse of dimensionality. The value of sliced Wasserstein distance is the average of transportation cost between one-dimensional representations (projections) of original measures that are obtained by Radon Transform (RT). Despite its efficiency in the number of supports, estimating the sliced Wasserstein requires a relatively large number of projections in high-dimensional settings. Therefore, for applications where the number of supports is relatively small compared with the dimension, e.g., several deep learning applications where the mini-batch approaches are utilized, the complexities from matrix multiplication of Radon Transform become the main computational bottleneck. To address this issue, we propose to derive projections by linearly and randomly combining a smaller number of projections which are named bottleneck projections. We explain the usage of these projections by introducing Hierarchical Radon Transform (HRT) which is constructed by applying Radon Transform variants recursively. We then formulate the approach into a new metric between measures, named Hierarchical Sliced Wasserstein (HSW) distance. By proving the injectivity of HRT, we derive the metricity of HSW. Moreover, we investigate the theoretical properties of HSW including its connection to SW variants and its computational and sample complexities. Finally, we compare the computational cost and generative quality of HSW with the conventional SW on the task of deep generative modeling using various benchmark datasets including CIFAR10, CelebA, and Tiny ImageNet.
comment: Accepted to ICLR 2023, 29 pages, 8 figures, 3 tables,
♻ ☆ PandA: Unsupervised Learning of Parts and Appearances in the Feature Maps of GANs ICLR 2023
Recent advances in the understanding of Generative Adversarial Networks (GANs) have led to remarkable progress in visual editing and synthesis tasks, capitalizing on the rich semantics that are embedded in the latent spaces of pre-trained GANs. However, existing methods are often tailored to specific GAN architectures and are limited to either discovering global semantic directions that do not facilitate localized control, or require some form of supervision through manually provided regions or segmentation masks. In this light, we present an architecture-agnostic approach that jointly discovers factors representing spatial parts and their appearances in an entirely unsupervised fashion. These factors are obtained by applying a semi-nonnegative tensor factorization on the feature maps, which in turn enables context-aware local image editing with pixel-level control. In addition, we show that the discovered appearance factors correspond to saliency maps that localize concepts of interest, without using any labels. Experiments on a wide range of GAN architectures and datasets show that, in comparison to the state of the art, our method is far more efficient in terms of training time and, most importantly, provides much more accurate localized control. Our code is available at: https://github.com/james-oldfield/PandA.
comment: Accepted at ICLR 2023. Code available at: https://github.com/james-oldfield/PandA
♻ ☆ Modality-Agnostic Variational Compression of Implicit Neural Representations
We introduce a modality-agnostic neural compression algorithm based on a functional view of data and parameterised as an Implicit Neural Representation (INR). Bridging the gap between latent coding and sparsity, we obtain compact latent representations non-linearly mapped to a soft gating mechanism. This allows the specialisation of a shared INR network to each data item through subnetwork selection. After obtaining a dataset of such latent representations, we directly optimise the rate/distortion trade-off in a modality-agnostic space using neural compression. Variational Compression of Implicit Neural Representations (VC-INR) shows improved performance given the same representational capacity pre quantisation while also outperforming previous quantisation schemes used for other INR techniques. Our experiments demonstrate strong results over a large set of diverse modalities using the same algorithm without any modality-specific inductive biases. We show results on images, climate data, 3D shapes and scenes as well as audio and video, introducing VC-INR as the first INR-based method to outperform codecs as well-known and diverse as JPEG 2000, MP3 and AVC/HEVC on their respective modalities.
♻ ☆ On Best-Arm Identification with a Fixed Budget in Non-Parametric Multi-Armed Bandits
We lay the foundations of a non-parametric theory of best-arm identification in multi-armed bandits with a fixed budget T. We consider general, possibly non-parametric, models D for distributions over the arms; an overarching example is the model D = P(0,1) of all probability distributions over [0,1]. We propose upper bounds on the average log-probability of misidentifying the optimal arm based on information-theoretic quantities that correspond to infima over Kullback-Leibler divergences between some distributions in D and a given distribution. This is made possible by a refined analysis of the successive-rejects strategy of Audibert, Bubeck, and Munos (2010). We finally provide lower bounds on the same average log-probability, also in terms of the same new information-theoretic quantities; these lower bounds are larger when the (natural) assumptions on the considered strategies are stronger. All these new upper and lower bounds generalize existing bounds based, e.g., on gaps between distributions.
♻ ☆ Transformation-Based Models of Video Sequences
In this work we propose a simple unsupervised approach for next frame prediction in video. Instead of directly predicting the pixels in a frame given past frames, we predict the transformations needed for generating the next frame in a sequence, given the transformations of the past frames. This leads to sharper results, while using a smaller prediction model. In order to enable a fair comparison between different video frame prediction models, we also propose a new evaluation protocol. We use generated frames as input to a classifier trained with ground truth sequences. This criterion guarantees that models scoring high are those producing sequences which preserve discriminative features, as opposed to merely penalizing any deviation, plausible or not, from the ground truth. Our proposed approach compares favourably against more sophisticated ones on the UCF-101 data set, while also being more efficient in terms of the number of parameters and computational cost.
♻ ☆ DiGress: Discrete Denoising diffusion for graph generation ICLR 2023
This work introduces DiGress, a discrete denoising diffusion model for generating graphs with categorical node and edge attributes. Our model utilizes a discrete diffusion process that progressively edits graphs with noise, through the process of adding or removing edges and changing the categories. A graph transformer network is trained to revert this process, simplifying the problem of distribution learning over graphs into a sequence of node and edge classification tasks. We further improve sample quality by introducing a Markovian noise model that preserves the marginal distribution of node and edge types during diffusion, and by incorporating auxiliary graph-theoretic features. A procedure for conditioning the generation on graph-level features is also proposed. DiGress achieves state-of-the-art performance on molecular and non-molecular datasets, with up to 3x validity improvement on a planar graph dataset. It is also the first model to scale to the large GuacaMol dataset containing 1.3M drug-like molecules without the use of molecule-specific representations.
comment: 22 pages. Published as a conference paper at ICLR 2023
♻ ☆ Efficient Adaptive Activation Rounding for Post-Training Quantization
Post-training quantization (PTQ) attracts increasing attention due to its convenience in deploying quantized neural networks. Rounding is the primary source of quantization error, for which previous works adopt the rounding-to-nearest scheme with a constant border of 0.5. This work demonstrates that optimizing rounding schemes can improve model accuracy. By replacing the constant border with a simple border function, we can obtain the minimal error for multiplying two numbers and eliminate the bias of its expected value, which further benefits model accuracy. Based on this insight, we approximate the border function to make the incurred overhead negligible. We also jointly optimize propagated errors and global errors. We finally propose our AQuant framework, which can learn the border function automatically. Extensive experiments show that AQuant achieves noticeable improvements compared with state-of-the-art works and pushes the accuracy of ResNet-18 up to 60.31% under the 2-bit weight and activation post-training quantization.
♻ ☆ Model-Aware Contrastive Learning: Towards Escaping the Dilemmas
Contrastive learning (CL) continuously achieves significant breakthroughs across multiple domains. However, the most common InfoNCE based methods suffer from some existing dilemmas, e.g., uniformity-tolerance dilemma (UTD) and the gradient reduction. It has been identified that UTD can lead to unexpected performance degradation. We argue that the fixity of temperature is to blame for UTD. To tackle this challenge, we enrich the CL loss family by presenting a Model-Aware Contrastive Learning (MACL) strategy, whose temperature is adaptive to the magnitude of alignment that reflects the basic confidence of the instance discrimination task, then enables CL loss to adjust the penalty strength for hard negatives adaptively. Regarding another dilemma, the gradient reduction issue, we derive the limits of an involved gradient scaling factor, which allows us to explain from a unified perspective why some recent approaches are effective with fewer negative samples, and summarily present a gradient reweighting to escape this dilemma. Extensive remarkable empirical results in vision, sentence, and graph modality validate our approach's general improvement for representation learning and downstream tasks.
♻ ☆ Linear-Time Gromov Wasserstein Distances using Low Rank Couplings and Costs
The ability to align points across two related yet incomparable point clouds (e.g. living in different spaces) plays an important role in machine learning. The Gromov-Wasserstein (GW) framework provides an increasingly popular answer to such problems, by seeking a low-distortion, geometry-preserving assignment between these points. As a non-convex, quadratic generalization of optimal transport (OT), GW is NP-hard. While practitioners often resort to solving GW approximately as a nested sequence of entropy-regularized OT problems, the cubic complexity (in the number $n$ of samples) of that approach is a roadblock. We show in this work how a recent variant of the OT problem that restricts the set of admissible couplings to those having a low-rank factorization is remarkably well suited to the resolution of GW: when applied to GW, we show that this approach is not only able to compute a stationary point of the GW problem in time $O(n^2)$, but also uniquely positioned to benefit from the knowledge that the initial cost matrices are low-rank, to yield a linear time $O(n)$ GW approximation. Our approach yields similar results, yet orders of magnitude faster computation than the SoTA entropic GW approaches, on both simulated and real data.
♻ ☆ Flip Initial Features: Generalization of Neural Networks Under Sparse Features for Semi-supervised Node Classification
Graph neural networks (GNNs) have been widely used under semi-supervised settings. Prior studies have mainly focused on finding appropriate graph filters (e.g., aggregation schemes) to generalize well for both homophilic and heterophilic graphs. Even though these approaches are essential and effective, they still suffer from the sparsity in initial node features inherent in the bag-of-words representation. Common in semi-supervised learning where the training samples often fail to cover the entire dimensions of graph filters (hyperplanes), this can precipitate over-fitting of specific dimensions in the first projection matrix. To deal with this problem, we suggest a simple and novel strategy; create additional space by flipping the initial features and hyperplane simultaneously. Training in both the original and in the flip space can provide precise updates of learnable parameters. To the best of our knowledge, this is the first attempt that effectively moderates the overfitting problem in GNN. Extensive experiments on real-world datasets demonstrate that the proposed technique improves the node classification accuracy up to 40.2 %
♻ ☆ Learning to Shape Rewards using a Game of Two Partners
Reward shaping (RS) is a powerful method in reinforcement learning (RL) for overcoming the problem of sparse or uninformative rewards. However, RS typically relies on manually engineered shaping-reward functions whose construction is time-consuming and error-prone. It also requires domain knowledge which runs contrary to the goal of autonomous learning. We introduce Reinforcement Learning Optimising Shaping Algorithm (ROSA), an automated reward shaping framework in which the shaping-reward function is constructed in a Markov game between two agents. A reward-shaping agent (Shaper) uses switching controls to determine which states to add shaping rewards for more efficient learning while the other agent (Controller) learns the optimal policy for the task using these shaped rewards. We prove that ROSA, which adopts existing RL algorithms, learns to construct a shaping-reward function that is beneficial to the task thus ensuring efficient convergence to high performance policies. We demonstrate ROSA's properties in three didactic experiments and show its superior performance against state-of-the-art RS algorithms in challenging sparse reward environments.
♻ ☆ Efficient Human-in-the-loop System for Guiding DNNs Attention
Attention guidance is an approach to addressing dataset bias in deep learning, where the model relies on incorrect features to make decisions. Focusing on image classification tasks, we propose an efficient human-in-the-loop system to interactively direct the attention of classifiers to the regions specified by users, thereby reducing the influence of co-occurrence bias and improving the transferability and interpretability of a DNN. Previous approaches for attention guidance require the preparation of pixel-level annotations and are not designed as interactive systems. We present a new interactive method to allow users to annotate images with simple clicks, and study a novel active learning strategy to significantly reduce the number of annotations. We conducted both a numerical evaluation and a user study to evaluate the proposed system on multiple datasets. Compared to the existing non-active-learning approach which usually relies on huge amounts of polygon-based segmentation masks to fine-tune or train the DNNs, our system can save lots of labor and money and obtain a fine-tuned network that works better even when the dataset is biased. The experiment results indicate that the proposed system is efficient, reasonable, and reliable.
comment: 13 pages, 11 figures, proceeding of ACM IUI 2023, video https://youtu.be/2MD-z6vXKJ4
♻ ☆ Two-Stage Constrained Actor-Critic for Short Video Recommendation
The wide popularity of short videos on social media poses new opportunities and challenges to optimize recommender systems on the video-sharing platforms. Users sequentially interact with the system and provide complex and multi-faceted responses, including watch time and various types of interactions with multiple videos. One the one hand, the platforms aims at optimizing the users' cumulative watch time (main goal) in long term, which can be effectively optimized by Reinforcement Learning. On the other hand, the platforms also needs to satisfy the constraint of accommodating the responses of multiple user interactions (auxiliary goals) such like, follow, share etc. In this paper, we formulate the problem of short video recommendation as a Constrained Markov Decision Process (CMDP). We find that traditional constrained reinforcement learning algorithms can not work well in this setting. We propose a novel two-stage constrained actor-critic method: At stage one, we learn individual policies to optimize each auxiliary signal. At stage two, we learn a policy to (i) optimize the main signal and (ii) stay close to policies learned at the first stage, which effectively guarantees the performance of this main policy on the auxiliaries. Through extensive offline evaluations, we demonstrate effectiveness of our method over alternatives in both optimizing the main goal as well as balancing the others. We further show the advantage of our method in live experiments of short video recommendations, where it significantly outperforms other baselines in terms of both watch time and interactions. Our approach has been fully launched in the production system to optimize user experiences on the platform.
comment: arXiv admin note: substantial text overlap with arXiv:2205.13248
♻ ☆ Federated deep transfer learning for EEG decoding using multiple BCI tasks
Deep learning has been successful in BCI decoding. However, it is very data-hungry and requires pooling data from multiple sources. EEG data from various sources decrease the decoding performance due to negative transfer. Recently, transfer learning for EEG decoding has been suggested as a remedy and become subject to recent BCI competitions (e.g. BEETL), but there are two complications in combining data from many subjects. First, privacy is not protected as highly personal brain data needs to be shared (and copied across increasingly tight information governance boundaries). Moreover, BCI data are collected from different sources and are often based on different BCI tasks, which has been thought to limit their reusability. Here, we demonstrate a federated deep transfer learning technique, the Multi-dataset Federated Separate-Common-Separate Network (MF-SCSN) based on our previous work of SCSN, which integrates privacy-preserving properties into deep transfer learning to utilise data sets with different tasks. This framework trains a BCI decoder using different source data sets obtained from different imagery tasks (e.g. some data sets with hands and feet, vs others with single hands and tongue, etc). Therefore, by introducing privacy-preserving transfer learning techniques, we unlock the reusability and scalability of existing BCI data sets. We evaluated our federated transfer learning method on the NeurIPS 2021 BEETL competition BCI task. The proposed architecture outperformed the baseline decoder by 3%. Moreover, compared with the baseline and other transfer learning algorithms, our method protects the privacy of the brain data from different data centres.
comment: 4 pages, 3 figures
♻ ☆ Translate First Reorder Later: Leveraging Monotonicity in Semantic Parsing ACL
Prior work in semantic parsing has shown that conventional seq2seq models fail at compositional generalization tasks. This limitation led to a resurgence of methods that model alignments between sentences and their corresponding meaning representations, either implicitly through latent variables or explicitly by taking advantage of alignment annotations. We take the second direction and propose TPOL, a two-step approach that first translates input sentences monotonically and then reorders them to obtain the correct output. This is achieved with a modular framework comprising a Translator and a Reorderer component. We test our approach on two popular semantic parsing datasets. Our experiments show that by means of the monotonic translations, TPOL can learn reliable lexico-logical patterns from aligned data, significantly improving compositional generalization both over conventional seq2seq models, as well as over other approaches that exploit gold alignments.
comment: Accepted at Findings of ACL: EACL 2023. 8 pages, 4 figures, 4 tables
♻ ☆ Causal-Discovery Performance of ChatGPT in the context of Neuropathic Pain Diagnosis
ChatGPT has demonstrated exceptional proficiency in natural language conversation, e.g., it can answer a wide range of questions while no previous large language models can. Thus, we would like to push its limit and explore its ability to answer causal discovery questions by using a medical benchmark (Tu et al. 2019) in causal discovery.
♻ ☆ Fruit Ripeness Classification: a Survey
Fruit is a key crop in worldwide agriculture feeding millions of people. The standard supply chain of fruit products involves quality checks to guarantee freshness, taste, and, most of all, safety. An important factor that determines fruit quality is its stage of ripening. This is usually manually classified by experts in the field, which makes it a labor-intensive and error-prone process. Thus, there is an arising need for automation in the process of fruit ripeness classification. Many automatic methods have been proposed that employ a variety of feature descriptors for the food item to be graded. Machine learning and deep learning techniques dominate the top-performing methods. Furthermore, deep learning can operate on raw data and thus relieve the users from having to compute complex engineered features, which are often crop-specific. In this survey, we review the latest methods proposed in the literature to automatize fruit ripeness classification, highlighting the most common feature descriptors they operate on.
♻ ☆ A Theoretical Framework for AI Models Explainability
EXplainable Artificial Intelligence (XAI) is a vibrant research topic in the artificial intelligence community, with growing interest across methods and domains. Much has been written about the subject, yet XAI still lacks shared terminology and a framework capable of providing structural soundness to explanations. In our work, we address these issues by proposing a novel definition of explanation that is a synthesis of what can be found in the literature. We recognize that explanations are not atomic but the combination of evidence stemming from the model and its input-output mapping, and the human interpretation of this evidence. Furthermore, we fit explanations into the properties of faithfulness (i.e., the explanation being a true description of the model's inner workings and decision-making process) and plausibility (i.e., how much the explanation looks convincing to the user). Using our proposed theoretical framework simplifies how these properties are operationalized and it provides new insight into common explanation methods that we analyze as case studies.
♻ ☆ Direct Advantage Estimation NeurIPS 2022
The predominant approach in reinforcement learning is to assign credit to actions based on the expected return. However, we show that the return may depend on the policy in a way which could lead to excessive variance in value estimation and slow down learning. Instead, we show that the advantage function can be interpreted as causal effects and shares similar properties with causal representations. Based on this insight, we propose Direct Advantage Estimation (DAE), a novel method that can model the advantage function and estimate it directly from on-policy data while simultaneously minimizing the variance of the return without requiring the (action-)value function. We also relate our method to Temporal Difference methods by showing how value functions can be seamlessly integrated into DAE. The proposed method is easy to implement and can be readily adapted by modern actor-critic methods. We evaluate DAE empirically on three discrete control domains and show that it can outperform generalized advantage estimation (GAE), a strong baseline for advantage estimation, on a majority of the environments when applied to policy optimization.
comment: Published at NeurIPS 2022
♻ ☆ Human alignment of neural network representations ICLR 2023
Today's computer vision models achieve human or near-human level performance across a wide variety of vision tasks. However, their architectures, data, and learning algorithms differ in numerous ways from those that give rise to human vision. In this paper, we investigate the factors that affect the alignment between the representations learned by neural networks and human mental representations inferred from behavioral responses. We find that model scale and architecture have essentially no effect on the alignment with human behavioral responses, whereas the training dataset and objective function both have a much larger impact. These findings are consistent across three datasets of human similarity judgments collected using two different tasks. Linear transformations of neural network representations learned from behavioral responses from one dataset substantially improve alignment with human similarity judgments on the other two datasets. In addition, we find that some human concepts such as food and animals are well-represented by neural networks whereas others such as royal or sports-related objects are not. Overall, although models trained on larger, more diverse datasets achieve better alignment with humans than models trained on ImageNet alone, our results indicate that scaling alone is unlikely to be sufficient to train neural networks with conceptual representations that match those used by humans.
comment: Accepted for publication at ICLR 2023
♻ ☆ Re-parameterizing Your Optimizers rather than Architectures ICLR 2023
The well-designed structures in neural networks reflect the prior knowledge incorporated into the models. However, though different models have various priors, we are used to training them with model-agnostic optimizers such as SGD. In this paper, we propose to incorporate model-specific prior knowledge into optimizers by modifying the gradients according to a set of model-specific hyper-parameters. Such a methodology is referred to as Gradient Re-parameterization, and the optimizers are named RepOptimizers. For the extreme simplicity of model structure, we focus on a VGG-style plain model and showcase that such a simple model trained with a RepOptimizer, which is referred to as RepOpt-VGG, performs on par with or better than the recent well-designed models. From a practical perspective, RepOpt-VGG is a favorable base model because of its simple structure, high inference speed and training efficiency. Compared to Structural Re-parameterization, which adds priors into models via constructing extra training-time structures, RepOptimizers require no extra forward/backward computations and solve the problem of quantization. We hope to spark further research beyond the realms of model structure design. Code and models \url{https://github.com/DingXiaoH/RepOptimizers}.
comment: ICLR 2023
♻ ☆ GIPA: A General Information Propagation Algorithm for Graph Learning DASFAA2023
Graph neural networks (GNNs) have been widely used in graph-structured data computation, showing promising performance in various applications such as node classification, link prediction, and network recommendation. Existing works mainly focus on node-wise correlation when doing weighted aggregation of neighboring nodes based on attention, such as dot product by the dense vectors of two nodes. This may cause conflicting noise in nodes to be propagated when doing information propagation. To solve this problem, we propose a General Information Propagation Algorithm (GIPA in short), which exploits more fine-grained information fusion including bit-wise and feature-wise correlations based on edge features in their propagation. Specifically, the bit-wise correlation calculates the element-wise attention weight through a multi-layer perceptron (MLP) based on the dense representations of two nodes and their edge; The feature-wise correlation is based on the one-hot representations of node attribute features for feature selection. We evaluate the performance of GIPA on the Open Graph Benchmark proteins (OGBN-proteins for short) dataset and the Alipay dataset of Alibaba. Experimental results reveal that GIPA outperforms the state-of-the-art models in terms of prediction accuracy, e.g., GIPA achieves an average ROC-AUC of $0.8901\pm 0.0011$, which is better than that of all the existing methods listed in the OGBN-proteins leaderboard.
comment: Accepted by DASFAA2023. arXiv admin note: substantial text overlap with arXiv:2105.06035
♻ ☆ Show me your NFT and I tell you how it will perform: Multimodal representation learning for NFT selling price prediction
Non-Fungible Tokens (NFTs) represent deeds of ownership, based on blockchain technologies and smart contracts, of unique crypto assets on digital art forms (e.g., artworks or collectibles). In the spotlight after skyrocketing in 2021, NFTs have attracted the attention of crypto enthusiasts and investors intent on placing promising investments in this profitable market. However, the NFT financial performance prediction has not been widely explored to date. In this work, we address the above problem based on the hypothesis that NFT images and their textual descriptions are essential proxies to predict the NFT selling prices. To this purpose, we propose MERLIN, a novel multimodal deep learning framework designed to train Transformer-based language and visual models, along with graph neural network models, on collections of NFTs' images and texts. A key aspect in MERLIN is its independence on financial features, as it exploits only the primary data a user interested in NFT trading would like to deal with, i.e., NFT images and textual descriptions. By learning dense representations of such data, a price-category classification task is performed by MERLIN models, which can also be tuned according to user preferences in the inference phase to mimic different risk-return investment profiles. Experimental evaluation on a publicly available dataset has shown that MERLIN models achieve significant performances according to several financial assessment criteria, fostering profitable investments, and also beating baseline machine-learning classifiers based on financial features.
comment: Accepted paper at The ACM Web Conference 2023, April 30--May 04, 2023, Austin, Texas, USA
♻ ☆ StockEmotions: Discover Investor Emotions for Financial Sentiment Analysis and Multivariate Time Series AAAI-23
There has been growing interest in applying NLP techniques in the financial domain, however, resources are extremely limited. This paper introduces StockEmotions, a new dataset for detecting emotions in the stock market that consists of 10,000 English comments collected from StockTwits, a financial social media platform. Inspired by behavioral finance, it proposes 12 fine-grained emotion classes that span the roller coaster of investor emotion. Unlike existing financial sentiment datasets, StockEmotions presents granular features such as investor sentiment classes, fine-grained emotions, emojis, and time series data. To demonstrate the usability of the dataset, we perform a dataset analysis and conduct experimental downstream tasks. For financial sentiment/emotion classification tasks, DistilBERT outperforms other baselines, and for multivariate time series forecasting, a Temporal Attention LSTM model combining price index, text, and emotion features achieves the best performance than using a single feature.
comment: Preprint - Accepted by the AAAI-23 Bridge Program (AI for Financial Services)
♻ ☆ DCM: Deep energy method based on the principle of minimum complementary energy
The principle of minimum potential and complementary energy are the most important variational principles in solid mechanics. The deep energy method (DEM), which has received much attention, is based on the principle of minimum potential energy and lacks the important form of minimum complementary energy. Thus, we propose the deep energy method based on the principle of minimum complementary energy (DCM). The output function of DCM is the stress function that naturally satisfies the equilibrium equation. We extend the proposed DCM algorithm (DCM-P), adding the terms that naturally satisfy the biharmonic equation in the Airy stress function. We combine operator learning with physical equations and propose a deep complementary energy operator method (DCM-O), including branch net, trunk net, basis net, and particular net. DCM-O first combines existing high-fidelity numerical results to train DCM-O through data. Then the complementary energy is used to train the branch net and trunk net in DCM-O. To analyze DCM performance, we present the numerical result of the most common stress functions, the Prandtl and Airy stress function. The proposed method DCM is used to model the representative mechanical problems with the different types of boundary conditions. We compare DCM with the existing PINNs and DEM algorithms. The result shows the advantage of the proposed DCM is suitable for dealing with problems of dominated displacement boundary conditions, which is reflected in theory and our numerical experiments. DCM-P and DCM-O improve the accuracy of DCM and the speed of calculation convergence. DCM is an essential supplementary energy form of the deep energy method. We believe that operator learning based on the energy method can balance data and physical equations well, giving computational mechanics broad research prospects.
comment: 46 pages, 23 figures
♻ ☆ LAP: An Attention-Based Module for Faithful Interpretation and Knowledge Injection in Convolutional Neural Networks
Despite the state-of-the-art performance of deep convolutional neural networks, they are susceptible to bias and malfunction in unseen situations. The complex computation behind their reasoning is not sufficiently human-understandable to develop trust. External explainer methods have tried to interpret the network decisions in a human-understandable way, but they are accused of fallacies due to their assumptions and simplifications. On the other side, the inherent self-interpretability of models, while being more robust to the mentioned fallacies, cannot be applied to the already trained models. In this work, we propose a new attention-based pooling layer, called Local Attention Pooling (LAP), that accomplishes self-interpretability and the possibility for knowledge injection while improving the model's performance. Moreover, several weakly-supervised knowledge injection methodologies are provided to enhance the process of training. We verified our claims by evaluating several LAP-extended models on three different datasets, including Imagenet. The proposed framework offers more valid human-understandable and more faithful-to-the-model interpretations than the commonly used white-box explainer methods.
♻ ☆ Bayesian Optimization-based Combinatorial Assignment
We study the combinatorial assignment domain, which includes combinatorial auctions and course allocation. The main challenge in this domain is that the bundle space grows exponentially in the number of items. To address this, several papers have recently proposed machine learning-based preference elicitation algorithms that aim to elicit only the most important information from agents. However, the main shortcoming of this prior work is that it does not model a mechanism's uncertainty over values for not yet elicited bundles. In this paper, we address this shortcoming by presenting a Bayesian optimization-based combinatorial assignment (BOCA) mechanism. Our key technical contribution is to integrate a method for capturing model uncertainty into an iterative combinatorial auction mechanism. Concretely, we design a new method for estimating an upper uncertainty bound that can be used to define an acquisition function to determine the next query to the agents. This enables the mechanism to properly explore (and not just exploit) the bundle space during its preference elicitation phase. We run computational experiments in several spectrum auction domains to evaluate BOCA's performance. Our results show that BOCA achieves higher allocative efficiency than state-of-the-art approaches.
♻ ☆ Dueling RL: Reinforcement Learning with Trajectory Preferences
We consider the problem of preference based reinforcement learning (PbRL), where, unlike traditional reinforcement learning, an agent receives feedback only in terms of a 1 bit (0/1) preference over a trajectory pair instead of absolute rewards for them. The success of the traditional RL framework crucially relies on the underlying agent-reward model, which, however, depends on how accurately a system designer can express an appropriate reward function and often a non-trivial task. The main novelty of our framework is the ability to learn from preference-based trajectory feedback that eliminates the need to hand-craft numeric reward models. This paper sets up a formal framework for the PbRL problem with non-markovian rewards, where the trajectory preferences are encoded by a generalized linear model of dimension $d$. Assuming the transition model is known, we then propose an algorithm with almost optimal regret guarantee of $\tilde {\mathcal{O}}\left( SH d \log (T / \delta) \sqrt{T} \right)$. We further, extend the above algorithm to the case of unknown transition dynamics, and provide an algorithm with near optimal regret guarantee $\widetilde{\mathcal{O}}((\sqrt{d} + H^2 + |\mathcal{S}|)\sqrt{dT} +\sqrt{|\mathcal{S}||\mathcal{A}|TH} )$. To the best of our knowledge, our work is one of the first to give tight regret guarantees for preference based RL problems with trajectory preferences.
comment: Aadirupa Saha and Aldo Pacchiano contributed equally
♻ ☆ Predicting Molecule-Target Interaction by Learning Biomedical Network and Molecule Representations
The study of molecule-target interaction is quite important for drug discovery in terms of target identification, hit identification, pathway study, drug-drug interaction, etc. Most existing methodologies utilize either biomedical network information or molecule structural features to predict potential interaction link. However, the biomedical network information based methods usually suffer from cold start problem, while structure based methods often give limited performance due to the structure/interaction assumption and data quality. To address these issues, we propose a pseudo-siamese Graph Neural Network method, namely MTINet+, which learns both biomedical network topological and molecule structural/chemical information as representations to predict potential interaction of given molecule and target pair. In MTINet+, 1-hop subgraphs of given molecule and target pair are extracted from known interaction of biomedical network as topological information, meanwhile the molecule structural and chemical attributes are processed as molecule information. MTINet+ learns these two types of information as embedding features for predicting the pair link. In the experiments of different molecule-target interaction tasks, MTINet+ significantly outperforms over the state-of-the-art baselines. In addition, in our designed network sparsity experiments , MTINet+ shows strong robustness against different sparse biomedical networks.
comment: 9 pages, 6 figures. arXiv admin note: substantial text overlap with arXiv:2102.01649
♻ ☆ Dual PatchNorm
We propose Dual PatchNorm: two Layer Normalization layers (LayerNorms), before and after the patch embedding layer in Vision Transformers. We demonstrate that Dual PatchNorm outperforms the result of exhaustive search for alternative LayerNorm placement strategies in the Transformer block itself. In our experiments, incorporating this trivial modification, often leads to improved accuracy over well-tuned Vision Transformers and never hurts.
♻ ☆ Joint Reasoning on Hybrid-knowledge sources for Task-Oriented Dialog
Traditional systems designed for task oriented dialog utilize knowledge present only in structured knowledge sources to generate responses. However, relevant information required to generate responses may also reside in unstructured sources, such as documents. Recent state of the art models such as HyKnow and SeKnow aimed at overcoming these challenges make limiting assumptions about the knowledge sources. For instance, these systems assume that certain types of information, such as a phone number, is always present in a structured knowledge base (KB) while information about aspects such as entrance ticket prices, would always be available in documents. In this paper, we create a modified version of the MutliWOZ-based dataset prepared by SeKnow to demonstrate how current methods have significant degradation in performance when strict assumptions about the source of information are removed. Then, in line with recent work exploiting pre-trained language models, we fine-tune a BART based model using prompts for the tasks of querying knowledge sources, as well as, for response generation, without making assumptions about the information present in each knowledge source. Through a series of experiments, we demonstrate that our model is robust to perturbations to knowledge modality (source of information), and that it can fuse information from structured as well as unstructured knowledge to generate responses.
♻ ☆ Euclidean-Norm-Induced Schatten-p Quasi-Norm Regularization for Low-Rank Tensor Completion and Tensor Robust Principal Component Analysis
The nuclear norm and Schatten-$p$ quasi-norm are popular rank proxies in low-rank matrix recovery. However, computing the nuclear norm or Schatten-$p$ quasi-norm of a tensor is hard in both theory and practice, hindering their application to low-rank tensor completion (LRTC) and tensor robust principal component analysis (TRPCA). In this paper, we propose a new class of tensor rank regularizers based on the Euclidean norms of the CP component vectors of a tensor and show that these regularizers are monotonic transformations of tensor Schatten-$p$ quasi-norm. This connection enables us to minimize the Schatten-$p$ quasi-norm in LRTC and TRPCA implicitly via the component vectors. The method scales to big tensors and provides an arbitrarily sharper rank proxy for low-rank tensor recovery compared to the nuclear norm. On the other hand, we study the generalization abilities of LRTC with the Schatten-$p$ quasi-norm regularizer and LRTC with the proposed regularizers. The theorems show that a relatively sharper regularizer leads to a tighter error bound, which is consistent with our numerical results. Particularly, we prove that for LRTC with Schatten-$p$ quasi-norm regularizer on $d$-order tensors, $p=1/d$ is always better than any $p>1/d$ in terms of the generalization ability. We also provide a recovery error bound to verify the usefulness of small $p$ in the Schatten-$p$ quasi-norm for TRPCA. Numerical results on synthetic data and real data demonstrate the effectiveness of the regularization methods and theorems.
comment: 43 pages; published by Transactions on Machine Learning Research, January 2023; https://openreview.net/forum?id=Grhi800jVz
♻ ☆ Editing Language Model-based Knowledge Graph Embeddings
Recently decades have witnessed the empirical success of framing Knowledge Graph (KG) embeddings via language models. However, language model-based KG embeddings are usually deployed as static artifacts, which are challenging to modify without re-training after deployment. To address this issue, we propose a new task of editing language model-based KG embeddings in this paper. The proposed task aims to enable data-efficient and fast updates to KG embeddings without damaging the performance of the rest. We build four new datasets: E-FB15k237, A-FB15k237, E-WN18RR, and A-WN18RR, and evaluate several knowledge editing baselines demonstrating the limited ability of previous models to handle the proposed challenging task. We further propose a simple yet strong baseline dubbed KGEditor, which utilizes additional parametric layers of the hyper network to edit/add facts. Comprehensive experimental results demonstrate that KGEditor can perform better when updating specific facts while not affecting the rest with low training resources. Code and datasets will be available in https://github.com/zjunlp/PromptKG/tree/main/deltaKG.
comment: Work in progress and the project website is https://zjunlp.github.io/project/KGE_Editing/
♻ ☆ The Dual PC Algorithm and the Role of Gaussianity for Structure Learning of Bayesian Networks
Learning the graphical structure of Bayesian networks is key to describing data-generating mechanisms in many complex applications but poses considerable computational challenges. Observational data can only identify the equivalence class of the directed acyclic graph underlying a Bayesian network model, and a variety of methods exist to tackle the problem. Under certain assumptions, the popular PC algorithm can consistently recover the correct equivalence class by reverse-engineering the conditional independence (CI) relationships holding in the variable distribution. The dual PC algorithm is a novel scheme to carry out the CI tests within the PC algorithm by leveraging the inverse relationship between covariance and precision matrices. By exploiting block matrix inversions we can simultaneously perform tests on partial correlations of complementary (or dual) conditioning sets. The multiple CI tests of the dual PC algorithm proceed by first considering marginal and full-order CI relationships and progressively moving to central-order ones. Simulation studies show that the dual PC algorithm outperforms the classic PC algorithm both in terms of run time and in recovering the underlying network structure, even in the presence of deviations from Gaussianity. Additionally, we show that the dual PC algorithm applies for Gaussian copula models, and demonstrate its performance in that setting.
♻ ☆ Self-supervised Semi-implicit Graph Variational Auto-encoders with Masking
Generative graph self-supervised learning (SSL) aims to learn node representations by reconstructing the input graph data. However, most existing methods focus on unsupervised learning tasks only and very few work has shown its superiority over the state-of-the-art graph contrastive learning (GCL) models, especially on the classification task. While a very recent model has been proposed to bridge the gap, its performance on unsupervised learning tasks is still unknown. In this paper, to comprehensively enhance the performance of generative graph SSL against other GCL models on both unsupervised and supervised learning tasks, we propose the SeeGera model, which is based on the family of self-supervised variational graph auto-encoder (VGAE). Specifically, SeeGera adopts the semi-implicit variational inference framework, a hierarchical variational framework, and mainly focuses on feature reconstruction and structure/feature masking. On the one hand, SeeGera co-embeds both nodes and features in the encoder and reconstructs both links and features in the decoder. Since feature embeddings contain rich semantic information on features, they can be combined with node embeddings to provide fine-grained knowledge for feature reconstruction. On the other hand, SeeGera adds an additional layer for structure/feature masking to the hierarchical variational framework, which boosts the model generalizability. We conduct extensive experiments comparing SeeGera with 9 other state-of-the-art competitors. Our results show that SeeGera can compare favorably against other state-of-the-art GCL methods in a variety of unsupervised and supervised learning tasks.
comment: Accepted by WebConf 2023
♻ ☆ Predicting the power grid frequency of European islands
Modelling, forecasting and overall understanding of the dynamics of the power grid and its frequency are essential for the safe operation of existing and future power grids. Much previous research was focused on large continental areas, while small systems, such as islands are less well-studied. These natural island systems are ideal testing environments for microgrid proposals and artificially islanded grid operation. In the present paper, we utilize measurements of the power grid frequency obtained in European islands: the Faroe Islands, Ireland, the Balearic Islands and Iceland and investigate how their frequency can be predicted, compared to the Nordic power system, acting as a reference. The Balearic islands are found to be particularly deterministic and easy to predict in contrast to hard-to-predict Iceland. Furthermore, we show that typically 2-4 weeks of data are needed to improve prediction performance beyond simple benchmarks.
comment: 17 pages
Boosting Exploration in Multi-Task Reinforcement Learning using Adversarial Networks
Advancements in reinforcement learning (RL) have been remarkable in recent years. However, the limitations of traditional training methods have become increasingly evident, particularly in meta-RL settings where agents face new, unseen tasks. Conventional training approaches are susceptible to failure in such situations as they need more robustness to adversity. Our proposed adversarial training regime for Multi-Task Reinforcement Learning (MT-RL) addresses the limitations of conventional training methods in RL, especially in meta-RL environments where the agent faces new tasks. The adversarial component challenges the agent, forcing it to improve its decision-making abilities in dynamic and unpredictable situations. This component operates without relying on manual intervention or domain-specific knowledge, making it a highly versatile solution. Experiments conducted in multiple MT-RL environments demonstrate that adversarial training leads to better exploration and a deeper understanding of the environment. The adversarial training regime for MT-RL presents a new perspective on training and development for RL agents and is a valuable contribution to the field.
♻ ☆ Multi-Domain Long-Tailed Learning by Augmenting Disentangled Representations
There is an inescapable long-tailed class-imbalance issue in many real-world classification problems. Existing long-tailed classification methods focus on the single-domain setting, where all examples are drawn from the same distribution. However, real-world scenarios often involve multiple domains with distinct imbalanced class distributions. We study this multi-domain long-tailed learning problem and aim to produce a model that generalizes well across all classes and domains. Towards that goal, we introduce TALLY, which produces invariant predictors by balanced augmenting hidden representations over domains and classes. Built upon a proposed selective balanced sampling strategy, TALLY achieves this by mixing the semantic representation of one example with the domain-associated nuisances of another, producing a new representation for use as data augmentation. To improve the disentanglement of semantic representations, TALLY further utilizes a domain-invariant class prototype that averages out domain-specific effects. We evaluate TALLY on four long-tailed variants of classical domain generalization benchmarks and two real-world imbalanced multi-domain datasets. The results indicate that TALLY consistently outperforms other state-of-the-art methods in both subpopulation shift and domain shift.
♻ ☆ Input Invex Neural Network
Connected decision boundaries are useful in several tasks like image segmentation, clustering, alpha-shape or defining a region in nD-space. However, the machine learning literature lacks methods for generating connected decision boundaries using neural networks. Thresholding an invex function, a generalization of a convex function, generates such decision boundaries. This paper presents two methods for constructing invex functions using neural networks. The first approach is based on constraining a neural network with Gradient Clipped-Gradient Penality (GCGP), where we clip and penalise the gradients. In contrast, the second one is based on the relationship of the invex function to the composition of invertible and convex functions. We employ connectedness as a basic interpretation method and create connected region-based classifiers. We show that multiple connected set based classifiers can approximate any classification function. In the experiments section, we use our methods for classification tasks using an ensemble of 1-vs-all models as well as using a single multiclass model on larger-scale datasets. The experiments show that connected set-based classifiers do not pose any disadvantage over ordinary neural network classifiers, but rather, enhance their interpretability. We also did an extensive study on the properties of invex function and connected sets for interpretability and network morphism with experiments on simulated and real-world data sets. Our study suggests that invex function is fundamental to understanding and applying locality and connectedness of input space which is useful for various downstream tasks.
comment: 42 pages, 23 figures
♻ ☆ Transformers as Algorithms: Generalization and Stability in In-context Learning
In-context learning (ICL) is a type of prompting where a transformer model operates on a sequence of (input, output) examples and performs inference on-the-fly. In this work, we formalize in-context learning as an algorithm learning problem where a transformer model implicitly constructs a hypothesis function at inference-time. We first explore the statistical aspects of this abstraction through the lens of multitask learning: We obtain generalization bounds for ICL when the input prompt is (1) a sequence of i.i.d. (input, label) pairs or (2) a trajectory arising from a dynamical system. The crux of our analysis is relating the excess risk to the stability of the algorithm implemented by the transformer. We characterize when transformer/attention architecture provably obeys the stability condition and also provide empirical verification. For generalization on unseen tasks, we identify an inductive bias phenomenon in which the transfer learning risk is governed by the task complexity and the number of MTL tasks in a highly predictable manner. Finally, we provide numerical evaluations that (1) demonstrate transformers can indeed implement near-optimal algorithms on classical regression problems with i.i.d. and dynamic data, (2) provide insights on stability, and (3) verify our theoretical predictions.
comment: Revised version significantly improves the stability guarantees and provides new experiments
♻ ☆ CCSL: A Causal Structure Learning Method from Multiple Unknown Environments
Most existing causal structure learning methods assume data collected from one environment and independent and identically distributed (i.i.d.). In some cases, data are collected from different subjects from multiple environments, which provides more information but might make the data non-identical or non-independent distribution. Some previous efforts try to learn causal structure from this type of data in two independent stages, i.e., first discovering i.i.d. groups from non-i.i.d. samples, then learning the causal structures from different groups. This straightforward solution ignores the intrinsic connections between the two stages, that is both the clustering stage and the learning stage should be guided by the same causal mechanism. Towards this end, we propose a unified Causal Cluster Structures Learning (named CCSL) method for causal discovery from non-i.i.d. data. This method simultaneously integrates the following two tasks: 1) clustering samples of the subjects with the same causal mechanism into different groups; 2) learning causal structures from the samples within the group. Specifically, for the former, we provide a Causality-related Chinese Restaurant Process to cluster samples based on the similarity of the causal structure; for the latter, we introduce a variational-inference-based approach to learn the causal structures. Theoretical results provide identification of the causal model and the clustering model under the linear non-Gaussian assumption. Experimental results on both simulated and real-world data further validate the correctness and effectiveness of the proposed method.
♻ ☆ Towards Understanding and Mitigating Dimensional Collapse in Heterogeneous Federated Learning ICLR 2023
Federated learning aims to train models collaboratively across different clients without the sharing of data for privacy considerations. However, one major challenge for this learning paradigm is the {\em data heterogeneity} problem, which refers to the discrepancies between the local data distributions among various clients. To tackle this problem, we first study how data heterogeneity affects the representations of the globally aggregated models. Interestingly, we find that heterogeneous data results in the global model suffering from severe {\em dimensional collapse}, in which representations tend to reside in a lower-dimensional space instead of the ambient space. Moreover, we observe a similar phenomenon on models locally trained on each client and deduce that the dimensional collapse on the global model is inherited from local models. In addition, we theoretically analyze the gradient flow dynamics to shed light on how data heterogeneity result in dimensional collapse for local models. To remedy this problem caused by the data heterogeneity, we propose {\sc FedDecorr}, a novel method that can effectively mitigate dimensional collapse in federated learning. Specifically, {\sc FedDecorr} applies a regularization term during local training that encourages different dimensions of representations to be uncorrelated. {\sc FedDecorr}, which is implementation-friendly and computationally-efficient, yields consistent improvements over baselines on standard benchmark datasets. Code: https://github.com/Yujun-Shi/FedCLS.
comment: camera ready version of ICLR 2023
♻ ☆ ERA-Solver: Error-Robust Adams Solver for Fast Sampling of Diffusion Probabilistic Models
Though denoising diffusion probabilistic models (DDPMs) have achieved remarkable generation results, the low sampling efficiency of DDPMs still limits further applications. Since DDPMs can be formulated as diffusion ordinary differential equations (ODEs), various fast sampling methods can be derived from solving diffusion ODEs. However, we notice that previous sampling methods with fixed analytical form are not robust with the error in the noise estimated from pretrained diffusion models. In this work, we construct an error-robust Adams solver (ERA-Solver), which utilizes the implicit Adams numerical method that consists of a predictor and a corrector. Different from the traditional predictor based on explicit Adams methods, we leverage a Lagrange interpolation function as the predictor, which is further enhanced with an error-robust strategy to adaptively select the Lagrange bases with lower error in the estimated noise. Experiments on Cifar10, LSUN-Church, and LSUN-Bedroom datasets demonstrate that our proposed ERA-Solver achieves 5.14, 9.42, and 9.69 Fenchel Inception Distance (FID) for image generation, with only 10 network evaluations.
comment: 16 pages, 12 figures
♻ ☆ Global Optimization with Parametric Function Approximation
We consider the problem of global optimization with noisy zeroth order oracles - a well-motivated problem useful for various applications ranging from hyper-parameter tuning for deep learning to new material design. Existing work relies on Gaussian processes or other non-parametric family, which suffers from the curse of dimensionality. In this paper, we propose a new algorithm GO-UCB that leverages a parametric family of functions (e.g., neural networks) instead. Under a realizable assumption and a few other mild geometric conditions, we show that GO-UCB achieves a cumulative regret of $\tilde{O}(\sqrt{T})$ where $T$ is the time horizon. At the core of GO-UCB is a carefully designed uncertainty set over parameters based on gradients that allows optimistic exploration. Synthetic and real-world experiments illustrate GO-UCB works better than Bayesian optimization approaches in high dimensional cases, even if the model is misspecified.
♻ ☆ Introspective Experience Replay: Look Back When Surprised
In reinforcement learning (RL), experience replay-based sampling techniques play a crucial role in promoting convergence by eliminating spurious correlations. However, widely used methods such as uniform experience replay (UER) and prioritized experience replay (PER) have been shown to have sub-optimal convergence and high seed sensitivity respectively. To address these issues, we propose a novel approach called IntrospectiveExperience Replay (IER) that selectively samples batches of data points prior to surprising events. Our method builds upon the theoretically sound reverse experience replay (RER) technique, which has been shown to reduce bias in the output of Q-learning-type algorithms with linear function approximation. However, this approach is not always practical or reliable when using neural function approximation. Through empirical evaluations, we demonstrate that IER with neural function approximation yields reliable and superior performance compared toUER, PER, and hindsight experience replay (HER) across most tasks.
♻ ☆ Sparse GCA and Thresholded Gradient Descent
Generalized correlation analysis (GCA) is concerned with uncovering linear relationships across multiple datasets. It generalizes canonical correlation analysis that is designed for two datasets. We study sparse GCA when there are potentially multiple generalized correlation tuples in data and the loading matrix has a small number of nonzero rows. It includes sparse CCA and sparse PCA of correlation matrices as special cases. We first formulate sparse GCA as generalized eigenvalue problems at both population and sample levels via a careful choice of normalization constraints. Based on a Lagrangian form of the sample optimization problem, we propose a thresholded gradient descent algorithm for estimating GCA loading vectors and matrices in high dimensions. We derive tight estimation error bounds for estimators generated by the algorithm with proper initialization. We also demonstrate the prowess of the algorithm on a number of synthetic datasets.
Multimedia 2
☆ Fine-Grained Action Detection with RGB and Pose Information using Two Stream Convolutional Networks
As participants of the MediaEval 2022 Sport Task, we propose a two-stream network approach for the classification and detection of table tennis strokes. Each stream is a succession of 3D Convolutional Neural Network (CNN) blocks using attention mechanisms. Each stream processes different 4D inputs. Our method utilizes raw RGB data and pose information computed from MMPose toolbox. The pose information is treated as an image by applying the pose either on a black background or on the original RGB frame it has been computed from. Best performance is obtained by feeding raw RGB data to one stream, Pose + RGB (PRGB) information to the other stream and applying late fusion on the features. The approaches were evaluated on the provided TTStroke-21 data sets. We can report an improvement in stroke classification, reaching 87.3% of accuracy, while the detection does not outperform the baseline but still reaches an IoU of 0.349 and mAP of 0.110.
comment: Working note paper of the sport task of MediaEval 2022 in Bergen, Norway, 12-13 Jan 2023
☆ Baseline Method for the Sport Task of MediaEval 2022 with 3D CNNs using Attention Mechanisms
This paper presents the baseline method proposed for the Sports Video task part of the MediaEval 2022 benchmark. This task proposes two subtasks: stroke classification from trimmed videos, and stroke detection from untrimmed videos. This baseline addresses both subtasks. We propose two types of 3D-CNN architectures to solve the two subtasks. Both 3D-CNNs use Spatio-temporal convolutions and attention mechanisms. The architectures and the training process are tailored to solve the addressed subtask. This baseline method is shared publicly online to help the participants in their investigation and alleviate eventually some aspects of the task such as video processing, training method, evaluation and submission routine. The baseline method reaches 86.4% of accuracy with our v2 model for the classification subtask. For the detection subtask, the baseline reaches a mAP of 0.131 and IoU of 0.515 with our v1 model.
comment: Baseline paper for the sport Task of MediaEval 2022
Computation and Language 17
☆ TempEL: Linking Dynamically Evolving and Newly Emerging Entities
In our continuously evolving world, entities change over time and new, previously non-existing or unknown, entities appear. We study how this evolutionary scenario impacts the performance on a well established entity linking (EL) task. For that study, we introduce TempEL, an entity linking dataset that consists of time-stratified English Wikipedia snapshots from 2013 to 2022, from which we collect both anchor mentions of entities, and these target entities' descriptions. By capturing such temporal aspects, our newly introduced TempEL resource contrasts with currently existing entity linking datasets, which are composed of fixed mentions linked to a single static version of a target Knowledge Base (e.g., Wikipedia 2010 for CoNLL-AIDA). Indeed, for each of our collected temporal snapshots, TempEL contains links to entities that are continual, i.e., occur in all of the years, as well as completely new entities that appear for the first time at some point. Thus, we enable to quantify the performance of current state-of-the-art EL models for: (i) entities that are subject to changes over time in their Knowledge Base descriptions as well as their mentions' contexts, and (ii) newly created entities that were previously non-existing (e.g., at the time the EL model was trained). Our experimental results show that in terms of temporal performance degradation, (i) continual entities suffer a decrease of up to 3.1% EL accuracy, while (ii) for new entities this accuracy drop is up to 17.9%. This highlights the challenge of the introduced TempEL dataset and opens new research prospects in the area of time-evolving entity disambiguation.
☆ Hatemongers ride on echo chambers to escalate hate speech diffusion
Recent years have witnessed a swelling rise of hateful and abusive content over online social networks. While detection and moderation of hate speech have been the early go-to countermeasures, the solution requires a deeper exploration of the dynamics of hate generation and propagation. We analyze more than 32 million posts from over 6.8 million users across three popular online social networks to investigate the interrelations between hateful behavior, information dissemination, and polarised organization mediated by echo chambers. We find that hatemongers play a more crucial role in governing the spread of information compared to singled-out hateful content. This observation holds for both the growth of information cascades as well as the conglomeration of hateful actors. Dissection of the core-wise distribution of these networks points towards the fact that hateful users acquire a more well-connected position in the social network and often flock together to build up information cascades. We observe that this cohesion is far from mere organized behavior; instead, in these networks, hatemongers dominate the echo chambers -- groups of users actively align themselves to specific ideological positions. The observed dominance of hateful users to inflate information cascades is primarily via user interactions amplified within these echo chambers. We conclude our study with a cautionary note that popularity-based recommendation of content is susceptible to be exploited by hatemongers given their potential to escalate content popularity via echo-chambered interactions.
comment: Accepted in PNAS Nexus
☆ Nationality Bias in Text Generation EACL 2023
Little attention is placed on analyzing nationality bias in language models, especially when nationality is highly used as a factor in increasing the performance of social NLP models. This paper examines how a text generation model, GPT-2, accentuates pre-existing societal biases about country-based demonyms. We generate stories using GPT-2 for various nationalities and use sensitivity analysis to explore how the number of internet users and the country's economic status impacts the sentiment of the stories. To reduce the propagation of biases through large language models (LLM), we explore the debiasing method of adversarial triggering. Our results show that GPT-2 demonstrates significant bias against countries with lower internet users, and adversarial triggering effectively reduces the same.
comment: EACL 2023 Paper
☆ FineDeb: A Debiasing Framework for Language Models AAAI 2023
As language models are increasingly included in human-facing machine learning tools, bias against demographic subgroups has gained attention. We propose FineDeb, a two-phase debiasing framework for language models that starts with contextual debiasing of embeddings learned by pretrained language models. The model is then fine-tuned on a language modeling objective. Our results show that FineDeb offers stronger debiasing in comparison to other methods which often result in models as biased as the original language model. Our framework is generalizable for demographics with multiple classes, and we demonstrate its effectiveness through extensive experiments and comparisons with state of the art techniques. We release our code and data on GitHub.
comment: Poster presentation at AAAI 2023: The Workshop on Artificial Intelligence for Social Good 2023 (https://amulyayadav.github.io/AI4SG2023/)
☆ deep learning of segment-level feature representation for speech emotion recognition in conversations
Accurately detecting emotions in conversation is a necessary yet challenging task due to the complexity of emotions and dynamics in dialogues. The emotional state of a speaker can be influenced by many different factors, such as interlocutor stimulus, dialogue scene, and topic. In this work, we propose a conversational speech emotion recognition method to deal with capturing attentive contextual dependency and speaker-sensitive interactions. First, we use a pretrained VGGish model to extract segment-based audio representation in individual utterances. Second, an attentive bi-directional gated recurrent unit (GRU) models contextual-sensitive information and explores intra- and inter-speaker dependencies jointly in a dynamic manner. The experiments conducted on the standard conversational dataset MELD demonstrate the effectiveness of the proposed method when compared against state-of the-art methods.
comment: 6 pages, 4 figures
☆ A Semantic Approach to Negation Detection and Word Disambiguation with Natural Language Processing
This study aims to demonstrate the methods for detecting negations in a sentence by uniquely evaluating the lexical structure of the text via word sense disambiguation. Additionally, the proposed method examined all the unique features of the related expressions within a text to resolve the contextual usage of the sentence and the effect of negation on sentiment analysis. The application of popular expression detectors skips this important step, thereby neglecting the root words caught in the web of negation, and making text classification difficult for machine learning and sentiment analysis. This study adopts the Natural Language Processing (NLP) approach to discover and antonimize words that were negated for better accuracy in text classification. This method acts as a lens that reads through a given word sequence using a knowledge base provided by an NLP library called WordHoard in order to detect negation signals. Early results show that our initial analysis improved traditional sentiment analysis that sometimes neglects word negations or assigns an inverse polarity score. The SentiWordNet analyzer was improved by 35%, the Vader analyzer by 20% and the TextBlob analyzer by 6%.
☆ Unleashing the True Potential of Sequence-to-Sequence Models for Sequence Tagging and Structure Parsing ACL 2023
Sequence-to-Sequence (S2S) models have achieved remarkable success on various text generation tasks. However, learning complex structures with S2S models remains challenging as external neural modules and additional lexicons are often supplemented to predict non-textual outputs. We present a systematic study of S2S modeling using contained decoding on four core tasks: part-of-speech tagging, named entity recognition, constituency and dependency parsing, to develop efficient exploitation methods costing zero extra parameters. In particular, 3 lexically diverse linearization schemas and corresponding constrained decoding methods are designed and evaluated. Experiments show that although more lexicalized schemas yield longer output sequences that require heavier training, their sequences being closer to natural language makes them easier to learn. Moreover, S2S models using our constrained decoding outperform other S2S approaches using external resources. Our best models perform better than or comparably to the state-of-the-art for all 4 tasks, lighting a promise for S2S models to generate non-sequential structures.
comment: Accepted to TACL 2023: Transactions of the Association for Computational Linguistics, post-acceptance final version
☆ Exploring Data Augmentation for Code Generation Tasks EACL 2023
Advances in natural language processing, such as transfer learning from pre-trained language models, have impacted how models are trained for programming language tasks too. Previous research primarily explored code pre-training and expanded it through multi-modality and multi-tasking, yet the data for downstream tasks remain modest in size. Focusing on data utilization for downstream tasks, we propose and adapt augmentation methods that yield consistent improvements in code translation and summarization by up to 6.9% and 7.5% respectively. Further analysis suggests that our methods work orthogonally and show benefits in output code style and numeric consistency. We also discuss test data imperfections.
comment: Findings of EACL 2023
☆ PAMP: A unified framework boosting low resource automatic speech recognition
We propose a novel text-to-speech (TTS) data augmentation framework for low resource automatic speech recognition (ASR) tasks, named phoneme audio mix up (PAMP). The PAMP method is highly interpretable and can incorporate prior knowledge of pronunciation rules. Furthermore, PAMP can be easily deployed in almost any language, extremely for low resource ASR tasks. Extensive experiments have demonstrated the great effectiveness of PAMP on low resource ASR tasks: we achieve a \textbf{10.84\%} character error rate (CER) on the common voice Cantonese ASR task, bringing a great relative improvement of about \textbf{30\%} compared to the previous state-of-the-art which was achieved by fine-tuning the wav2vec2 pretrained model.
☆ Meta-Learning Siamese Network for Few-Shot Text Classification
Few-shot learning has been used to tackle the problem of label scarcity in text classification, of which meta-learning based methods have shown to be effective, such as the prototypical networks (PROTO). Despite the success of PROTO, there still exist three main problems: (1) ignore the randomness of the sampled support sets when computing prototype vectors; (2) disregard the importance of labeled samples; (3) construct meta-tasks in a purely random manner. In this paper, we propose a Meta-Learning Siamese Network, namely, Meta-SN, to address these issues. Specifically, instead of computing prototype vectors from the sampled support sets, Meta-SN utilizes external knowledge (e.g. class names and descriptive texts) for class labels, which is encoded as the low-dimensional embeddings of prototype vectors. In addition, Meta-SN presents a novel sampling strategy for constructing meta-tasks, which gives higher sampling probabilities to hard-to-classify samples. Extensive experiments are conducted on six benchmark datasets to show the clear superiority of Meta-SN over other state-of-the-art models. For reproducibility, all the datasets and codes are provided at https://github.com/hccngu/Meta-SN.
♻ ☆ A Non-monotonic Self-terminating Language Model ICLR 2023
Recent large-scale neural autoregressive sequence models have shown impressive performances on a variety of natural language generation tasks. However, their generated sequences often exhibit degenerate properties such as non-termination, undesirable repetition, and premature termination, when generated with decoding algorithms such as greedy search, beam search, top-$k$ sampling, and nucleus sampling. In this paper, we focus on the problem of non-terminating sequences resulting from an incomplete decoding algorithm. We first define an incomplete probable decoding algorithm which includes greedy search, top-$k$ sampling, and nucleus sampling, beyond the incomplete decoding algorithm originally put forward by Welleck et al. (2020). We then propose a non-monotonic self-terminating language model, which significantly relaxes the constraint of monotonically increasing termination probability in the originally proposed self-terminating language model by Welleck et al. (2020), to address the issue of non-terminating sequences when using incomplete probable decoding algorithms. We prove that our proposed model prevents non-terminating sequences when using not only incomplete probable decoding algorithms but also beam search. We empirically validate our model on sequence completion tasks with various architectures.
comment: Published as a conference paper at ICLR 2023
♻ ☆ EHRSQL: A Practical Text-to-SQL Benchmark for Electronic Health Records NeurIPS 2022
We present a new text-to-SQL dataset for electronic health records (EHRs). The utterances were collected from 222 hospital staff, including physicians, nurses, insurance review and health records teams, and more. To construct the QA dataset on structured EHR data, we conducted a poll at a university hospital and templatized the responses to create seed questions. Then, we manually linked them to two open-source EHR databases, MIMIC-III and eICU, and included them with various time expressions and held-out unanswerable questions in the dataset, which were all collected from the poll. Our dataset poses a unique set of challenges: the model needs to 1) generate SQL queries that reflect a wide range of needs in the hospital, including simple retrieval and complex operations such as calculating survival rate, 2) understand various time expressions to answer time-sensitive questions in healthcare, and 3) distinguish whether a given question is answerable or unanswerable based on the prediction confidence. We believe our dataset, EHRSQL, could serve as a practical benchmark to develop and assess QA models on structured EHR data and take one step further towards bridging the gap between text-to-SQL research and its real-life deployment in healthcare. EHRSQL is available at https://github.com/glee4810/EHRSQL.
comment: Published as a conference paper at NeurIPS 2022 (Track on Datasets and Benchmarks)
♻ ☆ Adaptive Perturbation-Based Gradient Estimation for Discrete Latent Variable Models AAAI
The integration of discrete algorithmic components in deep learning architectures has numerous applications. Recently, Implicit Maximum Likelihood Estimation (IMLE, Niepert, Minervini, and Franceschi 2021), a class of gradient estimators for discrete exponential family distributions, was proposed by combining implicit differentiation through perturbation with the path-wise gradient estimator. However, due to the finite difference approximation of the gradients, it is especially sensitive to the choice of the finite difference step size, which needs to be specified by the user. In this work, we present Adaptive IMLE (AIMLE), the first adaptive gradient estimator for complex discrete distributions: it adaptively identifies the target distribution for IMLE by trading off the density of gradient information with the degree of bias in the gradient estimates. We empirically evaluate our estimator on synthetic examples, as well as on Learning to Explain, Discrete Variational Auto-Encoders, and Neural Relational Inference tasks. In our experiments, we show that our adaptive gradient estimator can produce faithful estimates while requiring orders of magnitude fewer samples than other gradient estimators.
comment: Proceedings of the Thirty-Seventh AAAI Conference on Artificial Intelligence (AAAI 2023)
♻ ☆ MTEB: Massive Text Embedding Benchmark
Text embeddings are commonly evaluated on a small set of datasets from a single task not covering their possible applications to other tasks. It is unclear whether state-of-the-art embeddings on semantic textual similarity (STS) can be equally well applied to other tasks like clustering or reranking. This makes progress in the field difficult to track, as various models are constantly being proposed without proper evaluation. To solve this problem, we introduce the Massive Text Embedding Benchmark (MTEB). MTEB spans 8 embedding tasks covering a total of 58 datasets and 112 languages. Through the benchmarking of 33 models on MTEB, we establish the most comprehensive benchmark of text embeddings to date. We find that no particular text embedding method dominates across all tasks. This suggests that the field has yet to converge on a universal text embedding method and scale it up sufficiently to provide state-of-the-art results on all embedding tasks. MTEB comes with open-source code and a public leaderboard at https://github.com/embeddings-benchmark/mteb.
comment: 24 pages, 14 tables, 6 figures
♻ ☆ Learning to Speak from Text: Zero-Shot Multilingual Text-to-Speech with Unsupervised Text Pretraining
While neural text-to-speech (TTS) has achieved human-like natural synthetic speech, multilingual TTS systems are limited to resource-rich languages due to the need for paired text and studio-quality audio data. This paper proposes a method for zero-shot multilingual TTS using text-only data for the target language. The use of text-only data allows the development of TTS systems for low-resource languages for which only textual resources are available, making TTS accessible to thousands of languages. Inspired by the strong cross-lingual transferability of multilingual language models, our framework first performs masked language model pretraining with multilingual text-only data. Then we train this model with a paired data in a supervised manner, while freezing a language-aware embedding layer. This allows inference even for languages not included in the paired data but present in the text-only data. Evaluation results demonstrate highly intelligible zero-shot TTS with a character error rate of less than 12% for an unseen language. All experiments were conducted using public datasets and the implementation will be made available for reproducibility.
♻ ☆ Inverse scaling can become U-shaped
Scaling up language models has been empirically shown to improve performance and unlock emergent abilities. Conversely, observing worse performance as a function of scale ("inverse scaling") would indicate that scaling encourages behaviors that are misaligned with human preferences. The Inverse Scaling Prize identified eleven such inverse scaling tasks, evaluated on models of up to 280B parameters and up to 500 zettaFLOPs of training compute. This paper takes a closer look at these inverse scaling tasks. We evaluate models of up to 540B parameters, trained on five times more compute than those evaluated in the Inverse Scaling Prize. With this increased range of model sizes and training compute, ten out of the eleven tasks exhibit what we call "U-shaped scaling" -- performance decreases up to a certain model size, and then increases again up to the largest model evaluated. U-shaped scaling can be seen as emergent ability unlocked by scaling and implies that inverse scaling may not hold for larger models.
comment: v3 includes all 11 tasks in the inverse scaling benchmark
♻ ☆ Equi-Tuning: Group Equivariant Fine-Tuning of Pretrained Models
We introduce equi-tuning, a novel fine-tuning method that transforms (potentially non-equivariant) pretrained models into group equivariant models while incurring minimum $L_2$ loss between the feature representations of the pretrained and the equivariant models. Large pretrained models can be equi-tuned for different groups to satisfy the needs of various downstream tasks. Equi-tuned models benefit from both group equivariance as an inductive bias and semantic priors from pretrained models. We provide applications of equi-tuning on three different tasks: image classification, compositional generalization in language, and fairness in natural language generation (NLG). We also provide a novel group-theoretic definition for fairness in NLG. The effectiveness of this definition is shown by testing it against a standard empirical method of fairness in NLG. We provide experimental results for equi-tuning using a variety of pretrained models: Alexnet, Resnet, VGG, and Densenet for image classification; RNNs, GRUs, and LSTMs for compositional generalization; and GPT2 for fairness in NLG. We test these models on benchmark datasets across all considered tasks to show the generality and effectiveness of the proposed method.
Computer Vision and Pattern Recognition 43
☆ Reconstruction-driven motion estimation for motion-compensated MR CINE imaging
In cardiac CINE, motion-compensated MR reconstruction (MCMR) is an effective approach to address highly undersampled acquisitions by incorporating motion information between frames. In this work, we propose a deep learning-based framework to address the MCMR problem efficiently. Contrary to state-of-the-art (SOTA) MCMR methods which break the original problem into two sub-optimization problems, i.e. motion estimation and reconstruction, we formulate this problem as a single entity with one single optimization. We discard the canonical motion-warping loss (similarity measurement between motion-warped images and target images) to estimate the motion, but drive the motion estimation process directly by the final reconstruction performance. The higher reconstruction quality is achieved without using any smoothness loss terms and without iterative processing between motion estimation and reconstruction. Therefore, we avoid non-trivial loss weighting factors tuning and time-consuming iterative processing. Experiments on 43 in-house acquired 2D CINE datasets indicate that the proposed MCMR framework can deliver artifact-free motion estimation and high-quality MR images even for imaging accelerations up to 20x. The proposed framework is compared to SOTA non-MCMR and MCMR methods and outperforms these methods qualitatively and quantitatively in all applied metrics across all experiments with different acceleration rates.
☆ Leaving Reality to Imagination: Robust Classification via Generated Datasets
Recent research on robustness has revealed significant performance gaps between neural image classifiers trained on datasets that are similar to the test set, and those that are from a naturally shifted distribution, such as sketches, paintings, and animations of the object categories observed during training. Prior work focuses on reducing this gap by designing engineered augmentations of training data or through unsupervised pretraining of a single large model on massive in-the-wild training datasets scraped from the Internet. However, the notion of a dataset is also undergoing a paradigm shift in recent years. With drastic improvements in the quality, ease-of-use, and access to modern generative models, generated data is pervading the web. In this light, we study the question: How do these generated datasets influence the natural robustness of image classifiers? We find that Imagenet classifiers trained on real data augmented with generated data achieve higher accuracy and effective robustness than standard training and popular augmentation strategies in the presence of natural distribution shifts. We analyze various factors influencing these results, including the choice of conditioning strategies and the amount of generated data. Lastly, we introduce and analyze an evolving generated dataset, ImageNet-G-v1, to better benchmark the design, utility, and critique of standalone generated datasets for robust and trustworthy machine learning. The code and datasets are available at https://github.com/Hritikbansal/generative-robustness.
comment: 20 pages, 12 Figures, 8 Tables
☆ On the Role of Contrastive Representation Learning in Adversarial Robustness: An Empirical Study
Self-supervised contrastive learning has solved one of the significant obstacles in deep learning by alleviating the annotation cost. This advantage comes with the price of false negative-pair selection without any label information. Supervised contrastive learning has emerged as an extension of contrastive learning to eliminate this issue. However, aside from accuracy, there is a lack of understanding about the impacts of adversarial training on the representations learned by these learning schemes. In this work, we utilize supervised learning as a baseline to comprehensively study the robustness of contrastive and supervised contrastive learning under different adversarial training scenarios. Then, we begin by looking at how adversarial training affects the learned representations in hidden layers, discovering more redundant representations between layers of the model. Our results on CIFAR-10 and CIFAR-100 image classification benchmarks demonstrate that this redundancy is highly reduced by adversarial fine-tuning applied to the contrastive learning scheme, leading to more robust representations. However, adversarial fine-tuning is not very effective for supervised contrastive learning and supervised learning schemes. Our code is released at https://github.com/softsys4ai/CL-Robustness.
☆ Handwriting and Drawing for Depression Detection: A Preliminary Study
The events of the past 2 years related to the pandemic have shown that it is increasingly important to find new tools to help mental health experts in diagnosing mood disorders. Leaving aside the longcovid cognitive (e.g., difficulty in concentration) and bodily (e.g., loss of smell) effects, the short-term covid effects on mental health were a significant increase in anxiety and depressive symptoms. The aim of this study is to use a new tool, the online handwriting and drawing analysis, to discriminate between healthy individuals and depressed patients. To this aim, patients with clinical depression (n = 14), individuals with high sub-clinical (diagnosed by a test rather than a doctor) depressive traits (n = 15) and healthy individuals (n = 20) were recruited and asked to perform four online drawing /handwriting tasks using a digitizing tablet and a special writing device. From the raw collected online data, seventeen drawing/writing features (categorized into five categories) were extracted, and compared among the three groups of the involved participants, through ANOVA repeated measures analyses. Results shows that Time features are more effective in discriminating between healthy and participants with sub-clinical depressive characteristics. On the other hand, Ductus and Pressure features are more effective in discriminating between clinical depressed and healthy participants.
comment: In: Mahmud, M., Ieracitano, C., Kaiser, M.S., Mammone, N., Morabito, F.C. (eds) Applied Intelligence and Informatics. AII 2022. Communications in Computer and Information Science, vol 1724. Springer, Cham
☆ Multi-Task Self-Supervised Learning for Image Segmentation Task
Thanks to breakthroughs in AI and Deep learning methodology, Computer vision techniques are rapidly improving. Most computer vision applications require sophisticated image segmentation to comprehend what is image and to make an analysis of each section easier. Training deep learning networks for semantic segmentation required a large amount of annotated data, which presents a major challenge in practice as it is expensive and labor-intensive to produce such data. The paper presents 1. Self-supervised techniques to boost semantic segmentation performance using multi-task learning with Depth prediction and Surface Normalization . 2. Performance evaluation of the different types of weighing techniques (UW, Nash-MTL) used for Multi-task learning. NY2D dataset was used for performance evaluation. According to our evaluation, the Nash-MTL method outperforms single task learning(Semantic Segmentation).
☆ Deep Learning Approach for Early Stage Lung Cancer Detection
Lung cancer is the leading cause of death among different types of cancers. Every year, the lives lost due to lung cancer exceed those lost to pancreatic, breast, and prostate cancer combined. The survival rate for lung cancer patients is very low compared to other cancer patients due to late diagnostics. Thus, early lung cancer diagnostics is crucial for patients to receive early treatments, increasing the survival rate or even becoming cancer-free. This paper proposed a deep-learning model for early lung cancer prediction and diagnosis from Computed Tomography (CT) scans. The proposed mode achieves high accuracy. In addition, it can be a beneficial tool to support radiologists' decisions in predicting and detecting lung cancer and its stage.
comment: Under review in FLAIRS 2023
☆ KDEformer: Accelerating Transformers via Kernel Density Estimation
Dot-product attention mechanism plays a crucial role in modern deep architectures (e.g., Transformer) for sequence modeling, however, na\"ive exact computation of this model incurs quadratic time and memory complexities in sequence length, hindering the training of long-sequence models. Critical bottlenecks are due to the computation of partition functions in the denominator of softmax function as well as the multiplication of the softmax matrix with the matrix of values. Our key observation is that the former can be reduced to a variant of the kernel density estimation (KDE) problem, and an efficient KDE solver can be further utilized to accelerate the latter via subsampling-based fast matrix products. Our proposed KDEformer can approximate the attention in sub-quadratic time with provable spectral norm bounds, while all prior results merely provide entry-wise error bounds. Empirically, we verify that KDEformer outperforms other attention approximations in terms of accuracy, memory, and runtime on various pre-trained models. On BigGAN image generation, we achieve better generative scores than the exact computation with over $4\times$ speedup. For ImageNet classification with T2T-ViT, KDEformer shows over $18\times$ speedup while the accuracy drop is less than $0.5\%$.
comment: 26 pages, 7 figures
☆ Spatio-Temporal Point Process for Multiple Object Tracking
Multiple Object Tracking (MOT) focuses on modeling the relationship of detected objects among consecutive frames and merge them into different trajectories. MOT remains a challenging task as noisy and confusing detection results often hinder the final performance. Furthermore, most existing research are focusing on improving detection algorithms and association strategies. As such, we propose a novel framework that can effectively predict and mask-out the noisy and confusing detection results before associating the objects into trajectories. In particular, we formulate such "bad" detection results as a sequence of events and adopt the spatio-temporal point process}to model such events. Traditionally, the occurrence rate in a point process is characterized by an explicitly defined intensity function, which depends on the prior knowledge of some specific tasks. Thus, designing a proper model is expensive and time-consuming, with also limited ability to generalize well. To tackle this problem, we adopt the convolutional recurrent neural network (conv-RNN) to instantiate the point process, where its intensity function is automatically modeled by the training data. Furthermore, we show that our method captures both temporal and spatial evolution, which is essential in modeling events for MOT. Experimental results demonstrate notable improvements in addressing noisy and confusing detection results in MOT datasets. An improved state-of-the-art performance is achieved by incorporating our baseline MOT algorithm with the spatio-temporal point process model.
comment: This manuscript is the accepted version for TNNLS(IEEE Transactions on Neural Networks and Learning Systems)
☆ Mixture of Diffusers for scene composition and high resolution image generation
Diffusion methods have been proven to be very effective to generate images while conditioning on a text prompt. However, and although the quality of the generated images is unprecedented, these methods seem to struggle when trying to generate specific image compositions. In this paper we present Mixture of Diffusers, an algorithm that builds over existing diffusion models to provide a more detailed control over composition. By harmonizing several diffusion processes acting on different regions of a canvas, it allows generating larger images, where the location of each object and style is controlled by a separate diffusion process.
☆ See You Soon: Decoupled Iterative Refinement Framework for Interacting Hands Reconstruction from a Single RGB Image
Reconstructing interacting hands from a single RGB image is a very challenging task. On the one hand, severe mutual occlusion and similar local appearance between two hands confuse the extraction of visual features, resulting in the misalignment of estimated hand meshes and the image. On the other hand, there are complex interaction patterns between interacting hands, which significantly increases the solution space of hand poses and increases the difficulty of network learning. In this paper, we propose a decoupled iterative refinement framework to achieve pixel-alignment hand reconstruction while efficiently modeling the spatial relationship between hands. Specifically, we define two feature spaces with different characteristics, namely 2D visual feature space and 3D joint feature space. First, we obtain joint-wise features from the visual feature map and utilize a graph convolution network and a transformer to perform intra- and inter-hand information interaction in the 3D joint feature space, respectively. Then, we project the joint features with global information back into the 2D visual feature space in an obfuscation-free manner and utilize the 2D convolution for pixel-wise enhancement. By performing multiple alternate enhancements in the two feature spaces, our method can achieve an accurate and robust reconstruction of interacting hands. Our method outperforms all existing two-hand reconstruction methods by a large margin on the InterHand2.6M dataset. Meanwhile, our method shows a strong generalization ability for in-the-wild images.
☆ Multi-View Masked World Models for Visual Robotic Manipulation
Visual robotic manipulation research and applications often use multiple cameras, or views, to better perceive the world. How else can we utilize the richness of multi-view data? In this paper, we investigate how to learn good representations with multi-view data and utilize them for visual robotic manipulation. Specifically, we train a multi-view masked autoencoder which reconstructs pixels of randomly masked viewpoints and then learn a world model operating on the representations from the autoencoder. We demonstrate the effectiveness of our method in a range of scenarios, including multi-view control and single-view control with auxiliary cameras for representation learning. We also show that the multi-view masked autoencoder trained with multiple randomized viewpoints enables training a policy with strong viewpoint randomization and transferring the policy to solve real-robot tasks without camera calibration and an adaptation procedure. Videos demonstrations in real-world experiments and source code are available at the project website: https://sites.google.com/view/mv-mwm.
comment: First two authors contributed equally. Project webpage: https://sites.google.com/view/mv-mwm
☆ Diffusion Model for Generative Image Denoising
In supervised learning for image denoising, usually the paired clean images and noisy images are collected or synthesised to train a denoising model. L2 norm loss or other distance functions are used as the objective function for training. It often leads to an over-smooth result with less image details. In this paper, we regard the denoising task as a problem of estimating the posterior distribution of clean images conditioned on noisy images. We apply the idea of diffusion model to realize generative image denoising. According to the noise model in denoising tasks, we redefine the diffusion process such that it is different from the original one. Hence, the sampling of the posterior distribution is a reverse process of dozens of steps from the noisy image. We consider three types of noise model, Gaussian, Gamma and Poisson noise. With the guarantee of theory, we derive a unified strategy for model training. Our method is verified through experiments on three types of noise models and achieves excellent performance.
☆ Eliminating Prior Bias for Semantic Image Editing via Dual-Cycle Diffusion
The recent success of text-to-image generation diffusion models has also revolutionized semantic image editing, enabling the manipulation of images based on query/target texts. Despite these advancements, a significant challenge lies in the potential introduction of prior bias in pre-trained models during image editing, e.g., making unexpected modifications to inappropriate regions. To this point, we present a novel Dual-Cycle Diffusion model that addresses the issue of prior bias by generating an unbiased mask as the guidance of image editing. The proposed model incorporates a Bias Elimination Cycle that consists of both a forward path and an inverted path, each featuring a Structural Consistency Cycle to ensure the preservation of image content during the editing process. The forward path utilizes the pre-trained model to produce the edited image, while the inverted path converts the result back to the source image. The unbiased mask is generated by comparing differences between the processed source image and the edited image to ensure that both conform to the same distribution. Our experiments demonstrate the effectiveness of the proposed method, as it significantly improves the D-CLIP score from 0.272 to 0.283. The code will be available at https://github.com/JohnDreamer/DualCycleDiffsion.
☆ ShiftDDPMs: Exploring Conditional Diffusion Models by Shifting Diffusion Trajectories AAAI 2023
Diffusion models have recently exhibited remarkable abilities to synthesize striking image samples since the introduction of denoising diffusion probabilistic models (DDPMs). Their key idea is to disrupt images into noise through a fixed forward process and learn its reverse process to generate samples from noise in a denoising way. For conditional DDPMs, most existing practices relate conditions only to the reverse process and fit it to the reversal of unconditional forward process. We find this will limit the condition modeling and generation in a small time window. In this paper, we propose a novel and flexible conditional diffusion model by introducing conditions into the forward process. We utilize extra latent space to allocate an exclusive diffusion trajectory for each condition based on some shifting rules, which will disperse condition modeling to all timesteps and improve the learning capacity of model. We formulate our method, which we call \textbf{ShiftDDPMs}, and provide a unified point of view on existing related methods. Extensive qualitative and quantitative experiments on image synthesis demonstrate the feasibility and effectiveness of ShiftDDPMs.
comment: Accepted by AAAI 2023 Conference
☆ FastPillars: A Deployment-friendly Pillar-based 3D Detector
The deployment of 3D detectors strikes one of the major challenges in real-world self-driving scenarios. Existing BEV-based (i.e., Bird Eye View) detectors favor sparse convolution (known as SPConv) to speed up training and inference, which puts a hard barrier for deployment especially for on-device applications. In this paper, we tackle the problem of efficient 3D object detection from LiDAR point clouds with deployment in mind. To reduce computational burden, we propose a pillar-based 3D detector with high performance from an industry perspective, termed FastPillars. Compared with previous methods, we introduce a more effective Max-and-Attention pillar encoding (MAPE) module, and redesigning a powerful and lightweight backbone CRVNet imbued with Cross Stage Partial network (CSP) in a reparameterization style, forming a compact feature representation framework. Extensive experiments demonstrate that our FastPillars surpasses the state-of-the-art 3D detectors regarding both on-device speed and performance. Specifically, FastPillars can be effectively deployed through TensorRT, obtaining real-time performance (24FPS) on a single RTX3070Ti GPU with 64.6 mAP on the nuScenes test set. Our code will be released.
☆ Towards Precision in Appearance-based Gaze Estimation in the Wild
Appearance-based gaze estimation systems have shown great progress recently, yet the performance of these techniques depend on the datasets used for training. Most of the existing gaze estimation datasets setup in interactive settings were recorded in laboratory conditions and those recorded in the wild conditions display limited head pose and illumination variations. Further, we observed little attention so far towards precision evaluations of existing gaze estimation approaches. In this work, we present a large gaze estimation dataset, PARKS-Gaze, with wider head pose and illumination variation and with multiple samples for a single Point of Gaze (PoG). The dataset contains 974 minutes of data from 28 participants with a head pose range of 60 degrees in both yaw and pitch directions. Our within-dataset and cross-dataset evaluations and precision evaluations indicate that the proposed dataset is more challenging and enable models to generalize on unseen participants better than the existing in-the-wild datasets. The project page can be accessed here: https://github.com/lrdmurthy/PARKS-Gaze
☆ Aggregation of Disentanglement: Reconsidering Domain Variations in Domain Generalization
Domain Generalization (DG) is a fundamental challenge for machine learning models, which aims to improve model generalization on various domains. Previous methods focus on generating domain invariant features from various source domains. However, we argue that the domain variantions also contain useful information, ie, classification-aware information, for downstream tasks, which has been largely ignored. Different from learning domain invariant features from source domains, we decouple the input images into Domain Expert Features and noise. The proposed domain expert features lie in a learned latent space where the images in each domain can be classified independently, enabling the implicit use of classification-aware domain variations. Based on the analysis, we proposed a novel paradigm called Domain Disentanglement Network (DDN) to disentangle the domain expert features from the source domain images and aggregate the source domain expert features for representing the target test domain. We also propound a new contrastive learning method to guide the domain expert features to form a more balanced and separable feature space. Experiments on the widely-used benchmarks of PACS, VLCS, OfficeHome, DomainNet, and TerraIncognita demonstrate the competitive performance of our method compared to the recently proposed alternatives.
☆ Explainable Machine Learning: The Importance of a System-Centric Perspective
The landscape in the context of several signal processing applications and even education appears to be significantly affected by the emergence of machine learning (ML) and in particular deep learning (DL).The main reason for this is the ability of DL to model complex and unknown relationships between signals and the tasks of interest. Particularly, supervised DL algorithms have been fairly successful at recognizing perceptually or semantically useful signal information in different applications. In all of these, the training process uses labeled data to learn a mapping function (typically implicitly) from signals to the desired information (class label or target label). The trained DL model is then expected to correctly recognize/classify relevant information in a given test signal. A DL based framework is therefore, in general, very appealing since the features and characteristics of the required mapping are learned almost exclusively from the data without resorting to explicit model/system development. The focus on implicit modeling however also raises the issue of lack of explainability/interpretability of the resultant DL based mapping or the black box problem. As a result, explainable ML/DL is an active research area where the primary goal is to elaborate how the ML/DL model arrived at a prediction. We however note that despite the efforts, the commentary on black box problem appears to lack a technical discussion from the view point of: a) its origin and underlying reasons, and b) its practical implications on the design and deployment of ML/DL systems. Accordingly, a reasonable question that can be raised is as follows. Can the traditional system-centric approach (which places emphasis on explicit system modeling) provide useful insights into the nature of black box problem, and help develop more transparent ML/DL systems?
☆ Using Intermediate Forward Iterates for Intermediate Generator Optimization
Score-based models have recently been introduced as a richer framework to model distributions in high dimensions and are generally more suitable for generative tasks. In score-based models, a generative task is formulated using a parametric model (such as a neural network) to directly learn the gradient of such high dimensional distributions, instead of the density functions themselves, as is done traditionally. From the mathematical point of view, such gradient information can be utilized in reverse by stochastic sampling to generate diverse samples. However, from a computational perspective, existing score-based models can be efficiently trained only if the forward or the corruption process can be computed in closed form. By using the relationship between the process and layers in a feed-forward network, we derive a backpropagation-based procedure which we call Intermediate Generator Optimization to utilize intermediate iterates of the process with negligible computational overhead. The main advantage of IGO is that it can be incorporated into any standard autoencoder pipeline for the generative task. We analyze the sample complexity properties of IGO to solve downstream tasks like Generative PCA. We show applications of the IGO on two dense predictive tasks viz., image extrapolation, and point cloud denoising. Our experiments indicate that obtaining an ensemble of generators for various time points is possible using first-order methods.
☆ Semi-Supervised Domain Adaptation with Source Label Adaptation
Semi-Supervised Domain Adaptation (SSDA) involves learning to classify unseen target data with a few labeled and lots of unlabeled target data, along with many labeled source data from a related domain. Current SSDA approaches usually aim at aligning the target data to the labeled source data with feature space mapping and pseudo-label assignments. Nevertheless, such a source-oriented model can sometimes align the target data to source data of the wrong classes, degrading the classification performance. This paper presents a novel source-adaptive paradigm that adapts the source data to match the target data. Our key idea is to view the source data as a noisily-labeled version of the ideal target data. Then, we propose an SSDA model that cleans up the label noise dynamically with the help of a robust cleaner component designed from the target perspective. Since the paradigm is very different from the core ideas behind existing SSDA approaches, our proposed model can be easily coupled with them to improve their performance. Empirical results on two state-of-the-art SSDA approaches demonstrate that the proposed model effectively cleans up the noise within the source labels and exhibits superior performance over those approaches across benchmark datasets.
☆ CIPER: Combining Invariant and Equivariant Representations Using Contrastive and Predictive Learning
Self-supervised representation learning (SSRL) methods have shown great success in computer vision. In recent studies, augmentation-based contrastive learning methods have been proposed for learning representations that are invariant or equivariant to pre-defined data augmentation operations. However, invariant or equivariant features favor only specific downstream tasks depending on the augmentations chosen. They may result in poor performance when a downstream task requires the counterpart of those features (e.g., when the task is to recognize hand-written digits while the model learns to be invariant to in-plane image rotations rendering it incapable of distinguishing "9" from "6"). This work introduces Contrastive Invariant and Predictive Equivariant Representation learning (CIPER). CIPER comprises both invariant and equivariant learning objectives using one shared encoder and two different output heads on top of the encoder. One output head is a projection head with a state-of-the-art contrastive objective to encourage invariance to augmentations. The other is a prediction head estimating the augmentation parameters, capturing equivariant features. Both heads are discarded after training and only the encoder is used for downstream tasks. We evaluate our method on static image tasks and time-augmented image datasets. Our results show that CIPER outperforms a baseline contrastive method on various tasks, especially when the downstream task requires the encoding of augmentation-related information.
comment: 10 pages, 4 figures, 3 tables
☆ Pyramid Self-attention Polymerization Learning for Semi-supervised Skeleton-based Action Recognition
Most semi-supervised skeleton-based action recognition approaches aim to learn the skeleton action representations only at the joint level, but neglect the crucial motion characteristics at the coarser-grained body (e.g., limb, trunk) level that provide rich additional semantic information, though the number of labeled data is limited. In this work, we propose a novel Pyramid Self-attention Polymerization Learning (dubbed as PSP Learning) framework to jointly learn body-level, part-level, and joint-level action representations of joint and motion data containing abundant and complementary semantic information via contrastive learning covering coarse-to-fine granularity. Specifically, to complement semantic information from coarse to fine granularity in skeleton actions, we design a new Pyramid Polymerizing Attention (PPA) mechanism that firstly calculates the body-level attention map, part-level attention map, and joint-level attention map, as well as polymerizes these attention maps in a level-by-level way (i.e., from body level to part level, and further to joint level). Moreover, we present a new Coarse-to-fine Contrastive Loss (CCL) including body-level contrast loss, part-level contrast loss, and joint-level contrast loss to jointly measure the similarity between the body/part/joint-level contrasting features of joint and motion data. Finally, extensive experiments are conducted on the NTU RGB+D and North-Western UCLA datasets to demonstrate the competitive performance of the proposed PSP Learning in the semi-supervised skeleton-based action recognition task. The source codes of PSP Learning are publicly available at https://github.com/1xbq1/PSP-Learning.
☆ Contrast with Reconstruct: Contrastive 3D Representation Learning Guided by Generative Pretraining
Mainstream 3D representation learning approaches are built upon contrastive or generative modeling pretext tasks, where great improvements in performance on various downstream tasks have been achieved. However, by investigating the methods of these two paradigms, we find that (i) contrastive models are data-hungry that suffer from a representation over-fitting issue; (ii) generative models have a data filling issue that shows inferior data scaling capacity compared to contrastive models. This motivates us to learn 3D representations by sharing the merits of both paradigms, which is non-trivial due to the pattern difference between the two paradigms. In this paper, we propose contrast with reconstruct (ReCon) that unifies these two paradigms. ReCon is trained to learn from both generative modeling teachers and cross-modal contrastive teachers through ensemble distillation, where the generative student guides the contrastive student. An encoder-decoder style ReCon-block is proposed that transfers knowledge through cross attention with stop-gradient, which avoids pretraining over-fitting and pattern difference issues. ReCon achieves a new state-of-the-art in 3D representation learning, e.g., 91.26% accuracy on ScanObjectNN. Codes will be released at https://github.com/qizekun/ReCon.
comment: Tech report
☆ Spatiotemporal Decouple-and-Squeeze Contrastive Learning for Semi-Supervised Skeleton-based Action Recognition
Contrastive learning has been successfully leveraged to learn action representations for addressing the problem of semi-supervised skeleton-based action recognition. However, most contrastive learning-based methods only contrast global features mixing spatiotemporal information, which confuses the spatial- and temporal-specific information reflecting different semantic at the frame level and joint level. Thus, we propose a novel Spatiotemporal Decouple-and-Squeeze Contrastive Learning (SDS-CL) framework to comprehensively learn more abundant representations of skeleton-based actions by jointly contrasting spatial-squeezing features, temporal-squeezing features, and global features. In SDS-CL, we design a new Spatiotemporal-decoupling Intra-Inter Attention (SIIA) mechanism to obtain the spatiotemporal-decoupling attentive features for capturing spatiotemporal specific information by calculating spatial- and temporal-decoupling intra-attention maps among joint/motion features, as well as spatial- and temporal-decoupling inter-attention maps between joint and motion features. Moreover, we present a new Spatial-squeezing Temporal-contrasting Loss (STL), a new Temporal-squeezing Spatial-contrasting Loss (TSL), and the Global-contrasting Loss (GL) to contrast the spatial-squeezing joint and motion features at the frame level, temporal-squeezing joint and motion features at the joint level, as well as global joint and motion features at the skeleton level. Extensive experimental results on four public datasets show that the proposed SDS-CL achieves performance gains compared with other competitive methods.
☆ CECT: Controllable Ensemble CNN and Transformer for COVID-19 image classification by capturing both local and global image features
Purpose: Most computer vision models are developed based on either convolutional neural network (CNN) or transformer, while the former (latter) method captures local (global) features. To relieve model performance limitations due to the lack of global (local) features, we develop a novel classification network named CECT by controllable ensemble CNN and transformer. Methods: The proposed CECT is composed of a CNN-based encoder block, a deconvolution-ensemble decoder block, and a transformer-based classification block. Different from conventional CNN- or transformer-based methods, our CECT can capture features at both multi-local and global scales, and the contribution of local features at different scales can be controlled with the proposed ensemble coefficients. Results: We evaluate CECT on two public COVID-19 datasets and it outperforms other state-of-the-art methods on all evaluation metrics. Conclusion: With remarkable feature capture ability, we believe CECT can also be used in other medical image classification scenarios to assist the diagnosis.
comment: 20 pages, 5 figures
☆ A Disparity Refinement Framework for Learning-based Stereo Matching Methods in Cross-domain Setting for Laparoscopic Images
Purpose: Stereo matching methods that enable depth estimation are crucial for visualization enhancement applications in computer-assisted surgery (CAS). Learning-based stereo matching methods are promising to predict accurate results on laparoscopic images. However, they require a large amount of training data, and their performance may be degraded due to domain shifts. Methods: Maintaining robustness and improving the accuracy of learning-based methods are still open problems. To overcome the limitations of learning-based methods, we propose a disparity refinement framework consisting of a local disparity refinement method and a global disparity refinement method to improve the results of learning-based stereo matching methods in a cross-domain setting. Those learning-based stereo matching methods are pre-trained on a large public dataset of natural images and are tested on two datasets of laparoscopic images. Results: Qualitative and quantitative results suggest that our proposed disparity framework can effectively refine disparity maps when they are noise-corrupted on an unseen dataset, without compromising prediction accuracy when the network can generalize well on an unseen dataset. Conclusion: Our proposed disparity refinement framework could work with learning-based methods to achieve robust and accurate disparity prediction. Yet, as a large laparoscopic dataset for training learning-based methods does not exist and the generalization ability of networks remains to be improved, the incorporation of the proposed disparity refinement framework into existing networks will contribute to improving their overall accuracy and robustness associated with depth estimation.
☆ Selecting the Best Optimizers for Deep Learning based Medical Image Segmentation
The goal of this work is to identify the best optimizers for deep learning in the context of cardiac image segmentation and to provide guidance on how to design segmentation networks with effective optimization strategies. Adaptive learning helps with fast convergence by starting with a larger learning rate (LR) and gradually decreasing it. Momentum optimizers are particularly effective at quickly optimizing neural networks within the accelerated schemes category. By revealing the potential interplay between these two types of algorithms (LR and momentum optimizers or momentum rate (MR) in short), in this article, we explore the two variants of SGD algorithms in a single setting. We suggest using cyclic learning as the base optimizer and integrating optimal values of learning rate and momentum rate. We investigated the relationship of LR and MR under an important problem of medical image segmentation of cardiac structures from MRI and CT scans. We conducted experiments using the cardiac imaging dataset from the ACDC challenge of MICCAI 2017, and four different architectures shown to be successful for cardiac image segmentation problems. Our comprehensive evaluations demonstrated that the proposed optimizer achieved better results (over a 2\% improvement in the dice metric) than other optimizers in deep learning literature with similar or lower computational cost in both single and multi-object segmentation settings. We hypothesized that combination of accelerated and adaptive optimization methods can have a drastic effect in medical image segmentation performances. To this end, we proposed a new cyclic optimization method (\textit{CLMR}) to address the efficiency and accuracy problems in deep learning based medical image segmentation. The proposed strategy yielded better generalization in comparison to adaptive optimizers.
☆ ReDi: Efficient Learning-Free Diffusion Inference via Trajectory Retrieval
Diffusion models show promising generation capability for a variety of data. Despite their high generation quality, the inference for diffusion models is still time-consuming due to the numerous sampling iterations required. To accelerate the inference, we propose ReDi, a simple yet learning-free Retrieval-based Diffusion sampling framework. From a precomputed knowledge base, ReDi retrieves a trajectory similar to the partially generated trajectory at an early stage of generation, skips a large portion of intermediate steps, and continues sampling from a later step in the retrieved trajectory. We theoretically prove that the generation performance of ReDi is guaranteed. Our experiments demonstrate that ReDi improves the model inference efficiency by 2x speedup. Furthermore, ReDi is able to generalize well in zero-shot cross-domain image generation such as image stylization.
☆ Design Booster: A Text-Guided Diffusion Model for Image Translation with Spatial Layout Preservation
Diffusion models are able to generate photorealistic images in arbitrary scenes. However, when applying diffusion models to image translation, there exists a trade-off between maintaining spatial structure and high-quality content. Besides, existing methods are mainly based on test-time optimization or fine-tuning model for each input image, which are extremely time-consuming for practical applications. To address these issues, we propose a new approach for flexible image translation by learning a layout-aware image condition together with a text condition. Specifically, our method co-encodes images and text into a new domain during the training phase. In the inference stage, we can choose images/text or both as the conditions for each time step, which gives users more flexible control over layout and content. Experimental comparisons of our method with state-of-the-art methods demonstrate our model performs best in both style image translation and semantic image translation and took the shortest time.
☆ Recurrence With Correlation Network for Medical Image Registration
We present Recurrence with Correlation Network (RWCNet), a medical image registration network with multi-scale features and a cost volume layer. We demonstrate that these architectural features improve medical image registration accuracy in two image registration datasets prepared for the MICCAI 2022 Learn2Reg Workshop Challenge. On the large-displacement National Lung Screening Test (NLST) dataset, RWCNet is able to achieve a total registration error (TRE) of 2.11mm between corresponding keypoints without instance fine-tuning. On the OASIS brain MRI dataset, RWCNet is able to achieve an average dice overlap of 81.7% for 35 different anatomical labels. It outperforms another multi-scale network, the Laplacian Image Registration Network (LapIRN), on both datasets. Ablation experiments are performed to highlight the contribution of the various architectural features. While multi-scale features improved validation accuracy for both datasets, the cost volume layer and number of recurrent steps only improved performance on the large-displacement NLST dataset. This result suggests that cost volume layer and iterative refinement using RNN provide good support for optimization and generalization in large-displacement medical image registration. The code for RWCNet is available at https://github.com/vigsivan/optimization-based-registration.
☆ JPEG Steganalysis Based on Steganographic Feature Enhancement and Graph Attention Learning
The purpose of image steganalysis is to determine whether the carrier image contains hidden information or not. Since JEPG is the most commonly used image format over social networks, steganalysis in JPEG images is also the most urgently needed to be explored. However, in order to detect whether secret information is hidden within JEPG images, the majority of existing algorithms are designed in conjunction with the popular computer vision related networks, without considering the key characteristics appeared in image steganalysis. It is crucial that the steganographic signal, as an extremely weak signal, can be enhanced during its representation learning process. Motivated by this insight, in this paper, we introduce a novel representation learning algorithm for JPEG steganalysis that is mainly consisting of a graph attention learning module and a feature enhancement module. The graph attention learning module is designed to avoid global feature loss caused by the local feature learning of convolutional neural network and reliance on depth stacking to extend the perceptual domain. The feature enhancement module is applied to prevent the stacking of convolutional layers from weakening the steganographic information. In addition, pretraining as a way to initialize the network weights with a large-scale dataset is utilized to enhance the ability of the network to extract discriminative features. We advocate pretraining with ALASKA2 for the model trained with BOSSBase+BOWS2. The experimental results indicate that the proposed algorithm outperforms previous arts in terms of detection accuracy, which has verified the superiority and applicability of the proposed work.
comment: https://scholar.google.com/citations?user=IdiF7M0AAAAJ&hl=en
☆ Divide and Compose with Score Based Generative Models
While score based generative models, or diffusion models, have found success in image synthesis, they are often coupled with text data or image label to be able to manipulate and conditionally generate images. Even though manipulation of images by changing the text prompt is possible, our understanding of the text embedding and our ability to modify it to edit images is quite limited. Towards the direction of having more control over image manipulation and conditional generation, we propose to learn image components in an unsupervised manner so that we can compose those components to generate and manipulate images in informed manner. Taking inspiration from energy based models, we interpret different score components as the gradient of different energy functions. We show how score based learning allows us to learn interesting components and we can visualize them through generation. We also show how this novel decomposition allows us to compose, generate and modify images in interesting ways akin to dreaming. We make our code available at https://github.com/sandeshgh/Score-based-disentanglement
♻ ☆ AbHE: All Attention-based Homography Estimation
Homography estimation is a basic computer vision task, which aims to obtain the transformation from multi-view images for image alignment. Unsupervised learning homography estimation trains a convolution neural network for feature extraction and transformation matrix regression. While the state-of-theart homography method is based on convolution neural networks, few work focuses on transformer which shows superiority in highlevel vision tasks. In this paper, we propose a strong-baseline model based on the Swin Transformer, which combines convolution neural network for local features and transformer module for global features. Moreover, a cross non-local layer is introduced to search the matched features within the feature maps coarsely. In the homography regression stage, we adopt an attention layer for the channels of correlation volume, which can drop out some weak correlation feature points. The experiment shows that in 8 Degree-of-Freedoms(DOFs) homography estimation our method overperforms the state-of-the-art method.
♻ ☆ RecolorNeRF: Layer Decomposed Radiance Fields for Efficient Color Editing of 3D Scenes
Radiance fields have gradually become a main representation of media. Although its appearance editing has been studied, how to achieve view-consistent recoloring in an efficient manner is still under explored. We present RecolorNeRF, a novel user-friendly color editing approach for the neural radiance fields. Our key idea is to decompose the scene into a set of pure-colored layers, forming a palette. By this means, color manipulation can be conducted by altering the color components of the palette directly. To support efficient palette-based editing, the color of each layer needs to be as representative as possible. In the end, the problem is formulated as an optimization problem, where the layers and their blending weights are jointly optimized with the NeRF itself. Extensive experiments show that our jointly-optimized layer decomposition can be used against multiple backbones and produce photo-realistic recolored novel-view renderings. We demonstrate that RecolorNeRF outperforms baseline methods both quantitatively and qualitatively for color editing even in complex real-world scenes.
♻ ☆ Efficient Distributed Vision Transformer Foundation Model for Medical Imaging through Random Masked Sampling
In spite of the recent success of deep learning in the medical domain, the problem of data scarcity in the medical domain gets aggravated due to privacy and data ownership issues. Distributed learning approaches including federated learning have been studied to alleviate the problems, but they suffer from cumbersome communication overheads and weakness in privacy protection. To address this, here we propose a self-supervised masked sampling distillation method for vision transformer that can be performed without continuous communication but still enhance privacy using a vision transformer-specific encryption method. The effectiveness of our method is demonstrated with extensive experiments on two medical domain data and two different downstream tasks, showing superior performances than those obtained with the existing distributed learning strategy as well as the fine-tuning only baseline. As the self-supervised model built with the proposed method is capable of having a general semantic understanding of the modality, we demonstrate its potential as a task-agnostic foundation model for various medical tasks, widening the applicability in the medical domain.
♻ ☆ Minimizing Trajectory Curvature of ODE-based Generative Models
Recent ODE/SDE-based generative models, such as diffusion models, rectified flows, and flow matching, define a generative process as a time reversal of a fixed forward process. Even though these models show impressive performance on large-scale datasets, numerical simulation requires multiple evaluations of a neural network, leading to a slow sampling speed. We attribute the reason to the high curvature of the learned generative trajectories, as it is directly related to the truncation error of a numerical solver. Based on the relationship between the forward process and the curvature, here we present an efficient method of training the forward process to minimize the curvature of generative trajectories without any ODE/SDE simulation. Experiments show that our method achieves a lower curvature than previous models and, therefore, decreased sampling costs while maintaining competitive performance. Code is available at https://github.com/sangyun884/fast-ode.
♻ ☆ SoMoFormer: Social-Aware Motion Transformer for Multi-Person Motion Prediction
Multi-person motion prediction remains a challenging problem, especially in the joint representation learning of individual motion and social interactions. Most prior methods only involve learning local pose dynamics for individual motion (without global body trajectory) and also struggle to capture complex interaction dependencies for social interactions. In this paper, we propose a novel Social-Aware Motion Transformer (SoMoFormer) to effectively model individual motion and social interactions in a joint manner. Specifically, SoMoFormer extracts motion features from sub-sequences in displacement trajectory space to effectively learn both local and global pose dynamics for each individual. In addition, we devise a novel social-aware motion attention mechanism in SoMoFormer to further optimize dynamics representations and capture interaction dependencies simultaneously via motion similarity calculation across time and social dimensions. On both short- and long-term horizons, we empirically evaluate our framework on multi-person motion datasets and demonstrate that our method greatly outperforms state-of-the-art methods of single- and multi-person motion prediction. Code will be made publicly available upon acceptance.
♻ ☆ Deep Instance Segmentation with Automotive Radar Detection Points
Automotive radar provides reliable environmental perception in all-weather conditions with affordable cost, but it hardly supplies semantic and geometry information due to the sparsity of radar detection points. With the development of automotive radar technologies in recent years, instance segmentation becomes possible by using automotive radar. Its data contain contexts such as radar cross section and micro-Doppler effects, and sometimes can provide detection when the field of view is obscured. The outcome from instance segmentation could be potentially used as the input of trackers for tracking targets. The existing methods often utilize a clustering-based classification framework, which fits the need of real-time processing but has limited performance due to minimum information provided by sparse radar detection points. In this paper, we propose an efficient method based on clustering of estimated semantic information to achieve instance segmentation for the sparse radar detection points. In addition, we show that the performance of the proposed approach can be further enhanced by incorporating the visual multi-layer perceptron. The effectiveness of the proposed method is verified by experimental results on the popular RadarScenes dataset, achieving 89.53% mean coverage and 86.97% mean average precision with the IoU threshold of 0.5, which is superior to other approaches in the literature. More significantly, the consumed memory is around 1MB, and the inference time is less than 40ms, indicating that our proposed algorithm is storage and time efficient. These two criteria ensure the practicality of the proposed method in real-world systems.
comment: 11 pages, 9 figures, 3 tables, accepted by IEEE Transactions on Intelligent Vehicles
♻ ☆ Contrastive Learning for Automotive mmWave Radar Detection Points Based Instance Segmentation SC 2022
The automotive mmWave radar plays a key role in advanced driver assistance systems (ADAS) and autonomous driving. Deep learning-based instance segmentation enables real-time object identification from the radar detection points. In the conventional training process, accurate annotation is the key. However, high-quality annotations of radar detection points are challenging to achieve due to their ambiguity and sparsity. To address this issue, we propose a contrastive learning approach for implementing radar detection points-based instance segmentation. We define the positive and negative samples according to the ground-truth label, apply the contrastive loss to train the model first, and then perform fine-tuning for the following downstream task. In addition, these two steps can be merged into one, and pseudo labels can be generated for the unlabeled data to improve the performance further. Thus, there are four different training settings for our method. Experiments show that when the ground-truth information is only available for a small proportion of the training data, our method still achieves a comparable performance to the approach trained in a supervised manner with 100% ground-truth information.
comment: Accepted by IEEE ITSC 2022
♻ ☆ Advances and Challenges in Multimodal Remote Sensing Image Registration
Over the past few decades, with the rapid development of global aerospace and aerial remote sensing technology, the types of sensors have evolved from the traditional monomodal sensors (e.g., optical sensors) to the new generation of multimodal sensors [e.g., multispectral, hyperspectral, light detection and ranging (LiDAR) and synthetic aperture radar (SAR) sensors]. These advanced devices can dynamically provide various and abundant multimodal remote sensing images with different spatial, temporal, and spectral resolutions according to different application requirements. Since then, it is of great scientific significance to carry out the research of multimodal remote sensing image registration, which is a crucial step for integrating the complementary information among multimodal data and making comprehensive observations and analysis of the Earths surface. In this work, we will present our own contributions to the field of multimodal image registration, summarize the advantages and limitations of existing multimodal image registration methods, and then discuss the remaining challenges and make a forward-looking prospect for the future development of the field.
comment: 10 pages, 4 figures
♻ ☆ Robust Fine-Tuning of Deep Neural Networks with Hessian-based Generalization Guarantees ICML 2022
We consider transfer learning approaches that fine-tune a pretrained deep neural network on a target task. We study the generalization properties of fine-tuning to understand the problem of overfitting, which commonly occurs in practice. Previous works have shown that constraining the distance from the initialization of fine-tuning improves generalization. Using a PAC-Bayesian analysis, we observe that besides distance from initialization, Hessians affect generalization through the noise stability of deep neural networks against noise injections. Motivated by the observation, we develop Hessian distance-based generalization bounds for a wide range of fine-tuning methods. Additionally, we study the robustness of fine-tuning in the presence of noisy labels. We design an algorithm incorporating consistent losses and distance-based regularization for fine-tuning, along with a generalization error guarantee under class conditional independent noise in the training set labels. We perform a detailed empirical study of our algorithm on various noisy environments and architectures. On six image classification tasks whose training labels are generated with programmatic labeling, we find a 3.26% accuracy gain over prior fine-tuning methods. Meanwhile, the Hessian distance measure of the fine-tuned model decreases by six times more than existing approaches.
comment: 36 pages, 5 figures, 8 tables (Fixed typos). ICML 2022
♻ ☆ PyMIC: A deep learning toolkit for annotation-efficient medical image segmentation
Background and Objective: Open-source deep learning toolkits are one of the driving forces for developing medical image segmentation models. Existing toolkits mainly focus on fully supervised segmentation and require full and accurate pixel-level annotations that are time-consuming and difficult to acquire for segmentation tasks, which makes learning from imperfect labels highly desired for reducing the annotation cost. We aim to develop a new deep learning toolkit to support annotation-efficient learning for medical image segmentation. Methods: Our proposed toolkit named PyMIC is a modular deep learning library for medical image segmentation tasks. In addition to basic components that support development of high-performance models for fully supervised segmentation, it contains several advanced components tailored for learning from imperfect annotations, such as loading annotated and unannounced images, loss functions for unannotated, partially or inaccurately annotated images, and training procedures for co-learning between multiple networks, etc. PyMIC supports development of semi-supervised, weakly supervised and noise-robust learning methods for medical image segmentation. Results: We present several illustrative medical image segmentation tasks based on PyMIC: (1) Achieving competitive performance on fully supervised learning; (2) Semi-supervised cardiac structure segmentation with only 10% training images annotated; (3) Weakly supervised segmentation using scribble annotations; and (4) Learning from noisy labels for chest radiograph segmentation. Conclusions: The PyMIC toolkit is easy to use and facilitates efficient development of medical image segmentation models with imperfect annotations. It is modular and flexible, which enables researchers to develop high-performance models with low annotation cost. The source code is available at: https://github.com/HiLab-git/PyMIC.
comment: 12 pages, 6 figures
♻ ☆ Weakly-guided Self-supervised Pretraining for Temporal Activity Detection AAAI 2023
Temporal Activity Detection aims to predict activity classes per frame, in contrast to video-level predictions in Activity Classification (i.e., Activity Recognition). Due to the expensive frame-level annotations required for detection, the scale of detection datasets is limited. Thus, commonly, previous work on temporal activity detection resorts to fine-tuning a classification model pretrained on large-scale classification datasets (e.g., Kinetics-400). However, such pretrained models are not ideal for downstream detection, due to the disparity between the pretraining and the downstream fine-tuning tasks. In this work, we propose a novel 'weakly-guided self-supervised' pretraining method for detection. We leverage weak labels (classification) to introduce a self-supervised pretext task (detection) by generating frame-level pseudo labels, multi-action frames, and action segments. Simply put, we design a detection task similar to downstream, on large-scale classification data, without extra annotations. We show that the models pretrained with the proposed weakly-guided self-supervised detection task outperform prior work on multiple challenging activity detection benchmarks, including Charades and MultiTHUMOS. Our extensive ablations further provide insights on when and how to use the proposed models for activity detection. Code is available at https://github.com/kkahatapitiya/SSDet.
comment: Published as a conference paper at AAAI 2023
Information Retrieval 9
☆ TWIN: TWo-stage Interest Network for Lifelong User Behavior Modeling in CTR Prediction at Kuaishou
Life-long user behavior modeling, i.e., extracting a user's hidden interests from rich historical behaviors in months or even years, plays a central role in modern CTR prediction systems. Conventional algorithms mostly follow two cascading stages: a simple General Search Unit (GSU) for fast and coarse search over tens of thousands of long-term behaviors and an Exact Search Unit (ESU) for effective Target Attention (TA) over the small number of finalists from GSU. Although efficient, existing algorithms mostly suffer from a crucial limitation: the \textit{inconsistent} target-behavior relevance metrics between GSU and ESU. As a result, their GSU usually misses highly relevant behaviors but retrieves ones considered irrelevant by ESU. In such case, the TA in ESU, no matter how attention is allocated, mostly deviates from the real user interests and thus degrades the overall CTR prediction accuracy. To address such inconsistency, we propose \textbf{TWo-stage Interest Network (TWIN)}, where our Consistency-Preserved GSU (CP-GSU) adopts the identical target-behavior relevance metric as the TA in ESU, making the two stages twins. Specifically, to break TA's computational bottleneck and extend it from ESU to GSU, or namely from behavior length $10^2$ to length $10^4-10^5$, we build a novel attention mechanism by behavior feature splitting. For the video inherent features of a behavior, we calculate their linear projection by efficient pre-computing \& caching strategies. And for the user-item cross features, we compress each into a one-dimentional bias term in the attention score calculation to save the computational cost. The consistency between two stages, together with the effective TA-based relevance metric in CP-GSU, contributes to significant performance gain in CTR prediction.
☆ Adversarial Learning Data Augmentation for Graph Contrastive Learning in Recommendation DASFAA2023
Recently, Graph Neural Networks (GNNs) achieve remarkable success in Recommendation. To reduce the influence of data sparsity, Graph Contrastive Learning (GCL) is adopted in GNN-based CF methods for enhancing performance. Most GCL methods consist of data augmentation and contrastive loss (e.g., InfoNCE). GCL methods construct the contrastive pairs by hand-crafted graph augmentations and maximize the agreement between different views of the same node compared to that of other nodes, which is known as the InfoMax principle. However, improper data augmentation will hinder the performance of GCL. InfoMin principle, that the good set of views shares minimal information and gives guidelines to design better data augmentation. In this paper, we first propose a new data augmentation (i.e., edge-operating including edge-adding and edge-dropping). Then, guided by InfoMin principle, we propose a novel theoretical guiding contrastive learning framework, named Learnable Data Augmentation for Graph Contrastive Learning (LDA-GCL). Our methods include data augmentation learning and graph contrastive learning, which follow the InfoMin and InfoMax principles, respectively. In implementation, our methods optimize the adversarial loss function to learn data augmentation and effective representations of users and items. Extensive experiments on four public benchmark datasets demonstrate the effectiveness of LDA-GCL.
comment: Accepted and to appear at DASFAA2023
☆ A Semantic Approach to Negation Detection and Word Disambiguation with Natural Language Processing
This study aims to demonstrate the methods for detecting negations in a sentence by uniquely evaluating the lexical structure of the text via word sense disambiguation. Additionally, the proposed method examined all the unique features of the related expressions within a text to resolve the contextual usage of the sentence and the effect of negation on sentiment analysis. The application of popular expression detectors skips this important step, thereby neglecting the root words caught in the web of negation, and making text classification difficult for machine learning and sentiment analysis. This study adopts the Natural Language Processing (NLP) approach to discover and antonimize words that were negated for better accuracy in text classification. This method acts as a lens that reads through a given word sequence using a knowledge base provided by an NLP library called WordHoard in order to detect negation signals. Early results show that our initial analysis improved traditional sentiment analysis that sometimes neglects word negations or assigns an inverse polarity score. The SentiWordNet analyzer was improved by 35%, the Vader analyzer by 20% and the TextBlob analyzer by 6%.
♻ ☆ MTEB: Massive Text Embedding Benchmark
Text embeddings are commonly evaluated on a small set of datasets from a single task not covering their possible applications to other tasks. It is unclear whether state-of-the-art embeddings on semantic textual similarity (STS) can be equally well applied to other tasks like clustering or reranking. This makes progress in the field difficult to track, as various models are constantly being proposed without proper evaluation. To solve this problem, we introduce the Massive Text Embedding Benchmark (MTEB). MTEB spans 8 embedding tasks covering a total of 58 datasets and 112 languages. Through the benchmarking of 33 models on MTEB, we establish the most comprehensive benchmark of text embeddings to date. We find that no particular text embedding method dominates across all tasks. This suggests that the field has yet to converge on a universal text embedding method and scale it up sufficiently to provide state-of-the-art results on all embedding tasks. MTEB comes with open-source code and a public leaderboard at https://github.com/embeddings-benchmark/mteb.
comment: 24 pages, 14 tables, 6 figures
♻ ☆ Layout-aware Webpage Quality Assessment
Identifying high-quality webpages is fundamental for real-world search engines, which can fulfil users' information need with the less cognitive burden. Early studies of \emph{webpage quality assessment} usually design hand-crafted features that may only work on particular categories of webpages (e.g., shopping websites, medical websites). They can hardly be applied to real-world search engines that serve trillions of webpages with various types and purposes. In this paper, we propose a novel layout-aware webpage quality assessment model currently deployed in our search engine. Intuitively, layout is a universal and critical dimension for the quality assessment of different categories of webpages. Based on this, we directly employ the meta-data that describes a webpage, i.e., Document Object Model (DOM) tree, as the input of our model. The DOM tree data unifies the representation of webpages with different categories and purposes and indicates the layout of webpages. To assess webpage quality from complex DOM tree data, we propose a graph neural network (GNN) based method that extracts rich layout-aware information that implies webpage quality in an end-to-end manner. Moreover, we improve the GNN method with an attentive readout function, external web categories and a category-aware sampling method. We conduct rigorous offline and online experiments to show that our proposed solution is effective in real search engines, improving the overall usability and user experience.
comment: 10 pages, 5 figures
♻ ☆ Modeling Adaptive Fine-grained Task Relatedness for Joint CTR-CVR Estimation
In modern advertising and recommender systems, multi-task learning (MTL) paradigm has been widely employed to jointly predict diverse user feedbacks (e.g. click and purchase). While, existing MTL approaches are either rigid to adapt to different scenarios, or only capture coarse-grained task relatedness, thus making it difficult to effectively transfer knowledge across tasks. To address these issues, in this paper, we propose an Adaptive Fine-grained Task Relatedness modeling approach, AdaFTR, for joint CTR-CVR estimation. Our approach is developed based on a parameter-sharing MTL architecture, and introduces a novel adaptive inter-task representation alignment method based on contrastive learning.Given an instance, the inter-task representations of the same instance are considered as positive, while the representations of another random instance are considered as negative. Furthermore, we explicitly model fine-grained task relatedness as the contrast strength (i.e. the temperature coefficient in InfoNCE loss) at the instance level. For this purpose, we build a relatedness prediction network, so that it can predict the contrast strength for inter-task representations of an instance. In this way, we can adaptively set the temperature for contrastive learning in a fine-grained way (i.e. instance level), so as to better capture task relatedness. Both offline evaluation with public e-commerce datasets and online test in a real advertising system at Alibaba have demonstrated the effectiveness of our approach.
comment: 10 pages, 4 figures
♻ ☆ PEPNet: Parameter and Embedding Personalized Network for Infusing with Personalized Prior Information
With the increase of content pages and display styles in online services such as online-shopping and video-watching websites, industrial-scale recommender systems face challenges in multi-domain and multi-task recommendations. The core of multi-task and multi-domain recommendation is to accurately capture user interests in different domains given different user behaviors. In this paper, we propose a plug-and-play \textit{\textbf{P}arameter and \textbf{E}mbedding \textbf{P}ersonalized \textbf{Net}work (\textbf{PEPNet})} for multi-task recommendation in the multi-domain setting. PEPNet takes features with strong biases as input and dynamically scales the bottom-layer embeddings and the top-layer DNN hidden units in the model through a gate mechanism. By mapping personalized priors to scaling weights ranging from 0 to 2, PEPNet introduces both parameter personalization and embedding personalization. Embedding Personalized Network (EPNet) selects and aligns embeddings with different semantics under multiple domains. Parameter Personalized Network (PPNet) influences DNN parameters to balance interdependent targets in multiple tasks. We have made a series of special engineering optimizations combining the Kuaishou training framework and the online deployment environment. We have deployed the model in Kuaishou apps, serving over 300 million daily users. Both online and offline experiments have demonstrated substantial improvements in multiple metrics. In particular, we have seen a more than 1\% online increase in three major scenarios.
♻ ☆ Correlative Preference Transfer with Hierarchical Hypergraph Network for Multi-Domain Recommendation WWW 2023
Advanced recommender systems usually involve multiple domains (such as scenarios or categories) for various marketing strategies, and users interact with them to satisfy diverse demands. The goal of multi-domain recommendation (MDR) is to improve the recommendation performance of all domains simultaneously. Conventional graph neural network based methods usually deal with each domain separately, or train a shared model to serve all domains. The former fails to leverage users' cross-domain behaviors, making the behavior sparseness issue a great obstacle. The latter learns shared user representation with respect to all domains, which neglects users' domain-specific preferences. In this paper we propose $\mathsf{H^3Trans}$, a hierarchical hypergraph network based correlative preference transfer framework for MDR, which represents multi-domain user-item interactions into a unified graph to help preference transfer. $\mathsf{H^3Trans}$ incorporates two hyperedge-based modules, namely dynamic item transfer (Hyper-I) and adaptive user aggregation (Hyper-U). Hyper-I extracts correlative information from multi-domain user-item feedbacks for eliminating domain discrepancy of item representations. Hyper-U aggregates users' scattered preferences in multiple domains and further exploits the high-order (not only pair-wise) connections to improve user representations. Experiments on both public and production datasets verify the superiority of $\mathsf{H^3Trans}$ for MDR.
comment: Accepted by WWW 2023 research track. The first two authors contributed equally
♻ ☆ Universal Vision-Language Dense Retrieval: Learning A Unified Representation Space for Multi-Modal Retrieval ICLR 2023
This paper presents Universal Vision-Language Dense Retrieval (UniVL-DR), which builds a unified model for multi-modal retrieval. UniVL-DR encodes queries and multi-modality resources in an embedding space for searching candidates from different modalities. To learn a unified embedding space for multi-modal retrieval, UniVL-DR proposes two techniques: 1) Universal embedding optimization strategy, which contrastively optimizes the embedding space using the modality-balanced hard negatives; 2) Image verbalization method, which bridges the modality gap between images and texts in the raw data space. UniVL-DR achieves the state-of-the-art on the multi-modal open-domain question answering benchmark, WebQA, and outperforms all retrieval models on the two subtasks, text-text retrieval and text-image retrieval. It demonstrates that universal multi-modal search is feasible to replace the divide-and-conquer pipeline with a united model and also benefits single/cross modality tasks. All source codes of this work are available at https://github.com/OpenMatch/UniVL-DR.
comment: Accepted by ICLR 2023
Multimedia 3
☆ Leaving Reality to Imagination: Robust Classification via Generated Datasets
Recent research on robustness has revealed significant performance gaps between neural image classifiers trained on datasets that are similar to the test set, and those that are from a naturally shifted distribution, such as sketches, paintings, and animations of the object categories observed during training. Prior work focuses on reducing this gap by designing engineered augmentations of training data or through unsupervised pretraining of a single large model on massive in-the-wild training datasets scraped from the Internet. However, the notion of a dataset is also undergoing a paradigm shift in recent years. With drastic improvements in the quality, ease-of-use, and access to modern generative models, generated data is pervading the web. In this light, we study the question: How do these generated datasets influence the natural robustness of image classifiers? We find that Imagenet classifiers trained on real data augmented with generated data achieve higher accuracy and effective robustness than standard training and popular augmentation strategies in the presence of natural distribution shifts. We analyze various factors influencing these results, including the choice of conditioning strategies and the amount of generated data. Lastly, we introduce and analyze an evolving generated dataset, ImageNet-G-v1, to better benchmark the design, utility, and critique of standalone generated datasets for robust and trustworthy machine learning. The code and datasets are available at https://github.com/Hritikbansal/generative-robustness.
comment: 20 pages, 12 Figures, 8 Tables
☆ OAcode: Overall Aesthetic 2D Barcode on Screen
Nowadays, two-dimensional (2D) barcodes have been widely used in various domains. And a series of aesthetic 2D barcode schemes have been proposed to improve the visual quality and readability of 2D barcodes for better integration with marketing materials. Yet we believe that the existing aesthetic 2D barcode schemes are partially aesthetic because they only beautify the data area but retain the position detection patterns with the blackwhite appearance of traditional 2D barcode schemes. Thus, in this paper, we propose the first overall aesthetic 2D barcode scheme, called OAcode, in which the position detection pattern is canceled. Its detection process is based on the pre-designed symmetrical data area of OAcode, whose symmetry could be used as the calibration signal to restore the perspective transformation in the barcode scanning process. Moreover, an enhanced demodulation method is proposed to resist the lens distortion common in the camera-shooting process. The experimental results illustrate that when 5$\times$5 cm OAcode is captured with a resolution of 720$\times$1280 pixels, at the screen-camera distance of 10 cm and the angle less or equal to 25{\deg}, OAcode has 100% detection rate and 99.5% demodulation accuracy. For 10$\times$10 cm OAcode, it could be extracted by consumer-grade mobile phones at a distance of 90 cm with around 90% accuracy.
comment: Published in: IEEE Transactions on Multimedia
☆ JPEG Steganalysis Based on Steganographic Feature Enhancement and Graph Attention Learning
The purpose of image steganalysis is to determine whether the carrier image contains hidden information or not. Since JEPG is the most commonly used image format over social networks, steganalysis in JPEG images is also the most urgently needed to be explored. However, in order to detect whether secret information is hidden within JEPG images, the majority of existing algorithms are designed in conjunction with the popular computer vision related networks, without considering the key characteristics appeared in image steganalysis. It is crucial that the steganographic signal, as an extremely weak signal, can be enhanced during its representation learning process. Motivated by this insight, in this paper, we introduce a novel representation learning algorithm for JPEG steganalysis that is mainly consisting of a graph attention learning module and a feature enhancement module. The graph attention learning module is designed to avoid global feature loss caused by the local feature learning of convolutional neural network and reliance on depth stacking to extend the perceptual domain. The feature enhancement module is applied to prevent the stacking of convolutional layers from weakening the steganographic information. In addition, pretraining as a way to initialize the network weights with a large-scale dataset is utilized to enhance the ability of the network to extract discriminative features. We advocate pretraining with ALASKA2 for the model trained with BOSSBase+BOWS2. The experimental results indicate that the proposed algorithm outperforms previous arts in terms of detection accuracy, which has verified the superiority and applicability of the proposed work.
comment: https://scholar.google.com/citations?user=IdiF7M0AAAAJ&hl=en
Computation and Language 25
☆ A Benchmark and Scoring Algorithm for Enriching Arabic Synonyms
This paper addresses the task of extending a given synset with additional synonyms taking into account synonymy strength as a fuzzy value. Given a mono/multilingual synset and a threshold (a fuzzy value [0-1]), our goal is to extract new synonyms above this threshold from existing lexicons. We present twofold contributions: an algorithm and a benchmark dataset. The dataset consists of 3K candidate synonyms for 500 synsets. Each candidate synonym is annotated with a fuzzy value by four linguists. The dataset is important for (i) understanding how much linguists (dis/)agree on synonymy, in addition to (ii) using the dataset as a baseline to evaluate our algorithm. Our proposed algorithm extracts synonyms from existing lexicons and computes a fuzzy value for each candidate. Our evaluations show that the algorithm behaves like a linguist and its fuzzy values are close to those proposed by linguists (using RMSE and MAE). The dataset and a demo page are publicly available at https://portal.sina.birzeit.edu/synonyms.
☆ Construction Grammar Provides Unique Insight into Neural Language Models
Construction Grammar (CxG) has recently been used as the basis for probing studies that have investigated the performance of large pretrained language models (PLMs) with respect to the structure and meaning of constructions. In this position paper, we make suggestions for the continuation and augmentation of this line of research. We look at probing methodology that was not designed with CxG in mind, as well as probing methodology that was designed for specific constructions. We analyse selected previous work in detail, and provide our view of the most important challenges and research questions that this promising new field faces.
comment: GURT 2023
☆ How Many and Which Training Points Would Need to be Removed to Flip this Prediction? EACL 2023
We consider the problem of identifying a minimal subset of training data $\mathcal{S}_t$ such that if the instances comprising $\mathcal{S}_t$ had been removed prior to training, the categorization of a given test point $x_t$ would have been different. Identifying such a set may be of interest for a few reasons. First, the cardinality of $\mathcal{S}_t$ provides a measure of robustness (if $|\mathcal{S}_t|$ is small for $x_t$, we might be less confident in the corresponding prediction), which we show is correlated with but complementary to predicted probabilities. Second, interrogation of $\mathcal{S}_t$ may provide a novel mechanism for contesting a particular model prediction: If one can make the case that the points in $\mathcal{S}_t$ are wrongly labeled or irrelevant, this may argue for overturning the associated prediction. Identifying $\mathcal{S}_t$ via brute-force is intractable. We propose comparatively fast approximation methods to find $\mathcal{S}_t$ based on influence functions, and find that -- for simple convex text classification models -- these approaches can often successfully identify relatively small sets of training examples which, if removed, would flip the prediction. To our knowledge, this is the first work in to investigate the problem of identifying a minimal training set necessary to flip a given prediction in the context of machine learning.
comment: Accepted to EACL 2023
☆ Invariants for neural automata
Computational modeling of neurodynamical systems often deploys neural networks and symbolic dynamics. A particular way for combining these approaches within a framework called vector symbolic architectures leads to neural automata. An interesting research direction we have pursued under this framework has been to consider mapping symbolic dynamics onto neurodynamics, represented as neural automata. This representation theory, enables us to ask questions, such as, how does the brain implement Turing computations. Specifically, in this representation theory, neural automata result from the assignment of symbols and symbol strings to numbers, known as G\"odel encoding. Under this assignment symbolic computation becomes represented by trajectories of state vectors in a real phase space, that allows for statistical correlation analyses with real-world measurements and experimental data. However, these assignments are usually completely arbitrary. Hence, it makes sense to address the problem question of, which aspects of the dynamics observed under such a representation is intrinsic to the dynamics and which are not. In this study, we develop a formally rigorous mathematical framework for the investigation of symmetries and invariants of neural automata under different encodings. As a central concept we define patterns of equality for such systems. We consider different macroscopic observables, such as the mean activation level of the neural network, and ask for their invariance properties. Our main result shows that only step functions that are defined over those patterns of equality are invariant under recodings, while the mean activation is not. Our work could be of substantial importance for related regression studies of real-world measurements with neurosymbolic processors for avoiding confounding results that are dependant on a particular encoding and not intrinsic to the dynamics.
comment: 28 pages, 8 figures
☆ LipFormer: Learning to Lipread Unseen Speakers based on Visual-Landmark Transformers
Lipreading refers to understanding and further translating the speech of a speaker in the video into natural language. State-of-the-art lipreading methods excel in interpreting overlap speakers, i.e., speakers appear in both training and inference sets. However, generalizing these methods to unseen speakers incurs catastrophic performance degradation due to the limited number of speakers in training bank and the evident visual variations caused by the shape/color of lips for different speakers. Therefore, merely depending on the visible changes of lips tends to cause model overfitting. To address this problem, we propose to use multi-modal features across visual and landmarks, which can describe the lip motion irrespective to the speaker identities. Then, we develop a sentence-level lipreading framework based on visual-landmark transformers, namely LipFormer. Specifically, LipFormer consists of a lip motion stream, a facial landmark stream, and a cross-modal fusion. The embeddings from the two streams are produced by self-attention, which are fed to the cross-attention module to achieve the alignment between visuals and landmarks. Finally, the resulting fused features can be decoded to output texts by a cascade seq2seq model. Experiments demonstrate that our method can effectively enhance the model generalization to unseen speakers.
comment: Under review
☆ Interaction Order Prediction for Temporal Graphs
Link prediction in graphs is a task that has been widely investigated. It has been applied in various domains such as knowledge graph completion, content/item recommendation, social network recommendations and so on. The initial focus of most research was on link prediction in static graphs. However, there has recently been abundant work on modeling temporal graphs, and consequently one of the tasks that has been researched is link prediction in temporal graphs. However, most of the existing work does not focus on the order of link formation, and only predicts the existence of links. In this study, we aim to predict the order of node interactions.
☆ Weight, Is Attention All We Need? AEIUOrder: Greedy Ordering of Layer Weight Matrices in Transformer Improves Translation
Prior work has attempted to understand the internal structures and functionalities of Transformer-based encoder-decoder architectures on the level of multi-head attention and feed-forward sublayers. Interpretations have focused on the encoder and decoder, along with the combinatorial possibilities of the self-attention, cross-attention, and feed-forward sublayers. Could we improve the quality of translation by diving into the Transformer sublayer abstractions and permuting its layer weight matrices? We propose AEIUOrder to greedily reorder layer weight matrices in the encoder by their well-trainedness, as measured by Random Matrix Theory (RMT) metrics, and reverse the ordering scheme for the encoder. The objective is to maximize Total well-trainedness in the encoder while the decoder structure serves to represent the reverse process of encoding. On the standard Transformer (6 layers, model dimension 512), AEIUOrder achieves a BLEU score of 34.62 (baseline 34.31) on the IWSLT 2016 German-to-English translation task, and 27.95 BLEU on the WMT 2014 English-to-German translation task (baseline 27.91). AEIUOrder is also realized on Transformers with various depths and embedding dimensions, showing significant improvements on deeper, wider models than on their shallower, slimmer counterparts. For instance, the 8-layer, 768-dimension and the 4-layer, 1024-dimension Transformers achieve respective 29.1 and 29.31 BLEU scores on the IWSLT 2016 English-to-German translation task (28.53 and 28.97 on respective baselines). Our results suggest that the RMT-motivated approach to maximize \textit{Total well-trainedness}, by greedily reordering its layer weight matrices, facilitates the model to learn representations and generate translations more effectively.
☆ A New cross-domain strategy based XAI models for fake news detection
In this study, we presented a four-level cross-domain strategy for fake news detection on pre-trained models. Cross-domain text classification is a task of a model adopting a target domain by using the knowledge of the source domain. Explainability is crucial in understanding the behaviour of these complex models. A fine-tune BERT model is used to. perform cross-domain classification with several experiments using datasets from different domains. Explanatory models like Anchor, ELI5, LIME and SHAP are used to design a novel explainable approach to cross-domain levels. The experimental analysis has given an ideal pair of XAI models on different levels of cross-domain.
☆ Knowledge Graph Completion Method Combined With Adaptive Enhanced Semantic Information
Translation models tend to ignore the rich semantic information in triads in the process of knowledge graph complementation. To remedy this shortcoming, this paper constructs a knowledge graph complementation method that incorporates adaptively enhanced semantic information. The hidden semantic information inherent in the triad is obtained by fine-tuning the BERT model, and the attention feature embedding method is used to calculate the semantic attention scores between relations and entities in positive and negative triads and incorporate them into the structural information to form a soft constraint rule for semantic information. The rule is added to the original translation model to realize the adaptive enhancement of semantic information. In addition, the method takes into account the effect of high-dimensional vectors on the effect, and uses the BERT-whitening method to reduce the dimensionality and generate a more efficient semantic vector representation. After experimental comparison, the proposed method performs better on both FB15K and WIN18 datasets, with a numerical improvement of about 2.6% compared with the original translation model, which verifies the reasonableness and effectiveness of the method.
☆ Theory of Mind May Have Spontaneously Emerged in Large Language Models
Theory of mind (ToM), or the ability to impute unobservable mental states to others, is central to human social interactions, communication, empathy, self-consciousness, and morality. We administer classic false-belief tasks, widely used to test ToM in humans, to several language models, without any examples or pre-training. Our results show that models published before 2022 show virtually no ability to solve ToM tasks. Yet, the January 2022 version of GPT-3 (davinci-002) solved 70% of ToM tasks, a performance comparable with that of seven-year-old children. Moreover, its November 2022 version (davinci-003), solved 93% of ToM tasks, a performance comparable with that of nine-year-old children. These findings suggest that ToM-like ability (thus far considered to be uniquely human) may have spontaneously emerged as a byproduct of language models' improving language skills.
☆ Improving Prediction Backward-Compatiblility in NLP Model Upgrade with Gated Fusion EACL 2023
When upgrading neural models to a newer version, new errors that were not encountered in the legacy version can be introduced, known as regression errors. This inconsistent behavior during model upgrade often outweighs the benefits of accuracy gain and hinders the adoption of new models. To mitigate regression errors from model upgrade, distillation and ensemble have proven to be viable solutions without significant compromise in performance. Despite the progress, these approaches attained an incremental reduction in regression which is still far from achieving backward-compatible model upgrade. In this work, we propose a novel method, Gated Fusion, that promotes backward compatibility via learning to mix predictions between old and new models. Empirical results on two distinct model upgrade scenarios show that our method reduces the number of regression errors by 62% on average, outperforming the strongest baseline by an average of 25%.
comment: Camera-ready for EACL 2023 Findings
☆ FGSI: Distant Supervision for Relation Extraction method based on Fine-Grained Semantic Information
The main purpose of relation extraction is to extract the semantic relationships between tagged pairs of entities in a sentence, which plays an important role in the semantic understanding of sentences and the construction of knowledge graphs. In this paper, we propose that the key semantic information within a sentence plays a key role in the relationship extraction of entities. We propose the hypothesis that the key semantic information inside the sentence plays a key role in entity relationship extraction. And based on this hypothesis, we split the sentence into three segments according to the location of the entity from the inside of the sentence, and find the fine-grained semantic features inside the sentence through the intra-sentence attention mechanism to reduce the interference of irrelevant noise information. The proposed relational extraction model can make full use of the available positive semantic information. The experimental results show that the proposed relation extraction model improves the accuracy-recall curves and P@N values compared with existing methods, which proves the effectiveness of this model.
☆ Heterogeneous Federated Knowledge Graph Embedding Learning and Unlearning WWW 2023
Federated Learning (FL) recently emerges as a paradigm to train a global machine learning model across distributed clients without sharing raw data. Knowledge Graph (KG) embedding represents KGs in a continuous vector space, serving as the backbone of many knowledge-driven applications. As a promising combination, federated KG embedding can fully take advantage of knowledge learned from different clients while preserving the privacy of local data. However, realistic problems such as data heterogeneity and knowledge forgetting still remain to be concerned. In this paper, we propose FedLU, a novel FL framework for heterogeneous KG embedding learning and unlearning. To cope with the drift between local optimization and global convergence caused by data heterogeneity, we propose mutual knowledge distillation to transfer local knowledge to global, and absorb global knowledge back. Moreover, we present an unlearning method based on cognitive neuroscience, which combines retroactive interference and passive decay to erase specific knowledge from local clients and propagate to the global model by reusing knowledge distillation. We construct new datasets for assessing realistic performance of the state-of-the-arts. Extensive experiments show that FedLU achieves superior results in both link prediction and knowledge forgetting.
comment: Accepted in the ACM Web Conference (WWW 2023)
☆ Lived Experience Matters: Automatic Detection of Stigma toward People Who Use Substances on Social Media
Stigma toward people who use substances (PWUS) is a leading barrier to seeking treatment. Further, those in treatment are more likely to drop out if they experience higher levels of stigmatization. While related concepts of hate speech and toxicity, including those targeted toward vulnerable populations, have been the focus of automatic content moderation research, stigma and, in particular, people who use substances have not. This paper explores stigma toward PWUS using a data set of roughly 5,000 public Reddit posts. We performed a crowd-sourced annotation task where workers are asked to annotate each post for the presence of stigma toward PWUS and answer a series of questions related to their experiences with substance use. Results show that workers who use substances or know someone with a substance use disorder are more likely to rate a post as stigmatizing. Building on this, we use a supervised machine learning framework that centers workers with lived substance use experience to label each Reddit post as stigmatizing. Modeling person-level demographics in addition to comment-level language results in a classification accuracy (as measured by AUC) of 0.69 -- a 17% increase over modeling language alone. Finally, we explore the linguist cues which distinguish stigmatizing content: PWUS substances and those who don't agree that language around othering ("people", "they") and terms like "addict" are stigmatizing, while PWUS (as opposed to those who do not) find discussions around specific substances more stigmatizing. Our findings offer insights into the nature of perceived stigma in substance use. Additionally, these results further establish the subjective nature of such machine learning tasks, highlighting the need for understanding their social contexts.
☆ Representation Deficiency in Masked Language Modeling
Masked Language Modeling (MLM) has been one of the most prominent approaches for pretraining bidirectional text encoders due to its simplicity and effectiveness. One notable concern about MLM is that the special $\texttt{[MASK]}$ symbol causes a discrepancy between pretraining data and downstream data as it is present only in pretraining but not in fine-tuning. In this work, we offer a new perspective on the consequence of such a discrepancy: We demonstrate empirically and theoretically that MLM pretraining allocates some model dimensions exclusively for representing $\texttt{[MASK]}$ tokens, resulting in a representation deficiency for real tokens and limiting the pretrained model's expressiveness when it is adapted to downstream data without $\texttt{[MASK]}$ tokens. Motivated by the identified issue, we propose MAE-LM, which pretrains the Masked Autoencoder architecture with MLM where $\texttt{[MASK]}$ tokens are excluded from the encoder. Empirically, we show that MAE-LM improves the utilization of model dimensions for real token representations, and MAE-LM consistently outperforms MLM-pretrained models across different pretraining settings and model sizes when fine-tuned on the GLUE and SQuAD benchmarks.
☆ Sentiment Analysis on YouTube Smart Phone Unboxing Video Reviews in Sri Lanka
Product-related reviews are based on users' experiences that are mostly shared on videos in YouTube. It is the second most popular website globally in 2021. People prefer to watch videos on recently released products prior to purchasing, in order to gather overall feedback and make worthy decisions. These videos are created by vloggers who are enthusiastic about technical materials and feedback is usually placed by experienced users of the product or its brand. Analyzing the sentiment of the user reviews gives useful insights into the product in general. This study is focused on three smartphone reviews, namely, Apple iPhone 13, Google Pixel 6, and Samsung Galaxy S21 which were released in 2021. VADER, which is a lexicon and rule-based sentiment analysis tool was used to classify each comment to its appropriate positive or negative orientation. All three smartphones show a positive sentiment from the users' perspective and iPhone 13 has the highest number of positive reviews. The resulting models have been tested using N\"aive Bayes, Decision Tree, and Support Vector Machine. Among these three classifiers, Support Vector Machine shows higher accuracies and F1-scores.
♻ ☆ RedHOT: A Corpus of Annotated Medical Questions, Experiences, and Claims on Social Media EACL
We present Reddit Health Online Talk (RedHOT), a corpus of 22,000 richly annotated social media posts from Reddit spanning 24 health conditions. Annotations include demarcations of spans corresponding to medical claims, personal experiences, and questions. We collect additional granular annotations on identified claims. Specifically, we mark snippets that describe patient Populations, Interventions, and Outcomes (PIO elements) within these. Using this corpus, we introduce the task of retrieving trustworthy evidence relevant to a given claim made on social media. We propose a new method to automatically derive (noisy) supervision for this task which we use to train a dense retrieval model; this outperforms baseline models. Manual evaluation of retrieval results performed by medical doctors indicate that while our system performance is promising, there is considerable room for improvement. Collected annotations (and scripts to assemble the dataset), are available at https://github.com/sominw/redhot.
comment: To appear in the 17th Conference of the European Chapter of the Association for Computational Linguistics (EACL), 2023
♻ ☆ Topic Modelling of Swedish Newspaper Articles about Coronavirus: a Case Study using Latent Dirichlet Allocation Method
Topic Modelling (TM) is from the research branches of natural language understanding (NLU) and natural language processing (NLP) that is to facilitate insightful analysis from large documents and datasets, such as a summarisation of main topics and the topic changes. This kind of discovery is getting more popular in real-life applications due to its impact on big data analytics. In this study, from the social-media and healthcare domain, we apply popular Latent Dirichlet Allocation (LDA) methods to model the topic changes in Swedish newspaper articles about Coronavirus. We describe the corpus we created including 6515 articles, methods applied, and statistics on topic changes over approximately 1 year and two months period of time from 17th January 2020 to 13th March 2021. We hope this work can be an asset for grounding applications of topic modelling and can be inspiring for similar case studies in an era with pandemics, to support socio-economic impact research as well as clinical and healthcare analytics. Our data and source code are openly available at https://github. com/poethan/Swed_Covid_TM Keywords: Latent Dirichlet Allocation (LDA); Topic Modelling; Coronavirus; Pandemics; Natural Language Understanding
comment: 14 pages, 13 figures
♻ ☆ Transfer Knowledge from Natural Language to Electrocardiography: Can We Detect Cardiovascular Disease Through Language Models? EACL 2023
Recent advancements in Large Language Models (LLMs) have drawn increasing attention since the learned embeddings pretrained on large-scale datasets have shown powerful ability in various downstream applications. However, whether the learned knowledge by LLMs can be transferred to clinical cardiology remains unknown. In this work, we aim to bridge this gap by transferring the knowledge of LLMs to clinical Electrocardiography (ECG). We propose an approach for cardiovascular disease diagnosis and automatic ECG diagnosis report generation. We also introduce an additional loss function by Optimal Transport (OT) to align the distribution between ECG and language embedding. The learned embeddings are evaluated on two downstream tasks: (1) automatic ECG diagnosis report generation, and (2) zero-shot cardiovascular disease detection. Our approach is able to generate high-quality cardiac diagnosis reports and also achieves competitive zero-shot classification performance even compared with supervised baselines, which proves the feasibility of transferring knowledge from LLMs to the cardiac domain.
comment: EACL 2023. arXiv admin note: text overlap with arXiv:2202.00567
♻ ☆ Bag-of-Vectors Autoencoders for Unsupervised Conditional Text Generation AACL 2022
Text autoencoders are often used for unsupervised conditional text generation by applying mappings in the latent space to change attributes to the desired values. Recently, Mai et al. (2020) proposed Emb2Emb, a method to learn these mappings in the embedding space of an autoencoder. However, their method is restricted to autoencoders with a single-vector embedding, which limits how much information can be retained. We address this issue by extending their method to Bag-of-Vectors Autoencoders (BoV-AEs), which encode the text into a variable-size bag of vectors that grows with the size of the text, as in attention-based models. This allows to encode and reconstruct much longer texts than standard autoencoders. Analogous to conventional autoencoders, we propose regularization techniques that facilitate learning meaningful operations in the latent space. Finally, we adapt Emb2Emb for a training scheme that learns to map an input bag to an output bag, including a novel loss function and neural architecture. Our empirical evaluations on unsupervised sentiment transfer show that our method performs substantially better than a standard autoencoder.
comment: Published at AACL 2022
♻ ☆ Active Learning for Multilingual Semantic Parser EACL 2023
Current multilingual semantic parsing (MSP) datasets are almost all collected by translating the utterances in the existing datasets from the resource-rich language to the target language. However, manual translation is costly. To reduce the translation effort, this paper proposes the first active learning procedure for MSP (AL-MSP). AL-MSP selects only a subset from the existing datasets to be translated. We also propose a novel selection method that prioritizes the examples diversifying the logical form structures with more lexical choices, and a novel hyperparameter tuning method that needs no extra annotation cost. Our experiments show that AL-MSP significantly reduces translation costs with ideal selection methods. Our selection method with proper hyperparameters yields better parsing performance than the other baselines on two multilingual datasets.
comment: EACL 2023 (findings)
♻ ☆ A Concept Knowledge Graph for User Next Intent Prediction at Alipay
This paper illustrates the technologies of user next intent prediction with a concept knowledge graph. The system has been deployed on the Web at Alipay, serving more than 100 million daily active users. To explicitly characterize user intent, we propose \textbf{AlipayKG}, which is an offline concept knowledge graph in the Life-Service domain modeling the historical behaviors of users, the rich content interacted by users and the relations between them. We further introduce a Transformer-based model which integrates expert rules from the knowledge graph to infer the online user's next intent. Experimental results demonstrate that the proposed system can effectively enhance the performance of the downstream tasks while retaining explainability.
comment: Work in progress
♻ ☆ DeepPSL: End-to-end perception and reasoning
We introduce DeepPSL a variant of probabilistic soft logic (PSL) to produce an end-to-end trainable system that integrates reasoning and perception. PSL represents first-order logic in terms of a convex graphical model -- hinge-loss Markov random fields (HL-MRFs). PSL stands out among probabilistic logic frameworks due to its tractability having been applied to systems of more than 1 billion ground rules. The key to our approach is to represent predicates in first-order logic using deep neural networks and then to approximately back-propagate through the HL-MRF and thus train every aspect of the first-order system being represented. We believe that this approach represents an interesting direction for the integration of deep learning and reasoning techniques with applications to knowledge base learning, multi-task learning, and explainability. Evaluation on three different tasks demonstrates that DeepPSL significantly outperforms state-of-the-art neuro-symbolic methods on scalability while achieving comparable or better accuracy.
♻ ☆ ImaginE: An Imagination-Based Automatic Evaluation Metric for Natural Language Generation EACL 2023
Automatic evaluations for natural language generation (NLG) conventionally rely on token-level or embedding-level comparisons with text references. This differs from human language processing, for which visual imagination often improves comprehension. In this work, we propose ImaginE, an imagination-based automatic evaluation metric for natural language generation. With the help of StableDiffusion, a state-of-the-art text-to-image generator, we automatically generate an image as the embodied imagination for the text snippet and compute the imagination similarity using contextual embeddings. Experiments spanning several text generation tasks demonstrate that adding machine-generated images with our ImaginE displays great potential in introducing multi-modal information into NLG evaluation, and improves existing automatic metrics' correlations with human similarity judgments in both reference-based and reference-free evaluation scenarios.
comment: EACL 2023
♻ ☆ Analogical Inference Enhanced Knowledge Graph Embedding AAAI 2023
Knowledge graph embedding (KGE), which maps entities and relations in a knowledge graph into continuous vector spaces, has achieved great success in predicting missing links in knowledge graphs. However, knowledge graphs often contain incomplete triples that are difficult to inductively infer by KGEs. To address this challenge, we resort to analogical inference and propose a novel and general self-supervised framework AnKGE to enhance KGE models with analogical inference capability. We propose an analogical object retriever that retrieves appropriate analogical objects from entity-level, relation-level, and triple-level. And in AnKGE, we train an analogy function for each level of analogical inference with the original element embedding from a well-trained KGE model as input, which outputs the analogical object embedding. In order to combine inductive inference capability from the original KGE model and analogical inference capability enhanced by AnKGE, we interpolate the analogy score with the base model score and introduce the adaptive weights in the score function for prediction. Through extensive experiments on FB15k-237 and WN18RR datasets, we show that AnKGE achieves competitive results on link prediction task and well performs analogical inference.
comment: Accepted by AAAI 2023
Computer Vision and Pattern Recognition 30
☆ CLiNet: Joint Detection of Road Network Centerlines in 2D and 3D
This work introduces a new approach for joint detection of centerlines based on image data by localizing the features jointly in 2D and 3D. In contrast to existing work that focuses on detection of visual cues, we explore feature extraction methods that are directly amenable to the urban driving task. To develop and evaluate our approach, a large urban driving dataset dubbed AV Breadcrumbs is automatically labeled by leveraging vector map representations and projective geometry to annotate over 900,000 images. Our results demonstrate potential for dynamic scene modeling across various urban driving scenarios. Our model achieves an F1 score of 0.684 and an average normalized depth error of 2.083. The code and data annotations are publicly available.
comment: 5 pages, 4 figures, 1 table. Under review at IEEE Intelligent Vehicles Symposium 2023
☆ Human-Imperceptible Identification with Learnable Lensless Imaging
Lensless imaging protects visual privacy by capturing heavily blurred images that are imperceptible for humans to recognize the subject but contain enough information for machines to infer information. Unfortunately, protecting visual privacy comes with a reduction in recognition accuracy and vice versa. We propose a learnable lensless imaging framework that protects visual privacy while maintaining recognition accuracy. To make captured images imperceptible to humans, we designed several loss functions based on total variation, invertibility, and the restricted isometry property. We studied the effect of privacy protection with blurriness on the identification of personal identity via a quantitative method based on a subjective evaluation. Moreover, we validate our simulation by implementing a hardware realization of lensless imaging with photo-lithographically printed masks.
Self-supervised Multi-view Disentanglement for Expansion of Visual Collections WSDM 2023
Image search engines enable the retrieval of images relevant to a query image. In this work, we consider the setting where a query for similar images is derived from a collection of images. For visual search, the similarity measurements may be made along multiple axes, or views, such as style and color. We assume access to a set of feature extractors, each of which computes representations for a specific view. Our objective is to design a retrieval algorithm that effectively combines similarities computed over representations from multiple views. To this end, we propose a self-supervised learning method for extracting disentangled view-specific representations for images such that the inter-view overlap is minimized. We show how this allows us to compute the intent of a collection as a distribution over views. We show how effective retrieval can be performed by prioritizing candidate expansion images that match the intent of a query collection. Finally, we present a new querying mechanism for image search enabled by composing multiple collections and perform retrieval under this setting using the techniques presented in this paper.
comment: A version of this paper has been accepted at WSDM 2023
☆ Revisiting Image Deblurring with an Efficient ConvNet
Image deblurring aims to recover the latent sharp image from its blurry counterpart and has a wide range of applications in computer vision. The Convolution Neural Networks (CNNs) have performed well in this domain for many years, and until recently an alternative network architecture, namely Transformer, has demonstrated even stronger performance. One can attribute its superiority to the multi-head self-attention (MHSA) mechanism, which offers a larger receptive field and better input content adaptability than CNNs. However, as MHSA demands high computational costs that grow quadratically with respect to the input resolution, it becomes impractical for high-resolution image deblurring tasks. In this work, we propose a unified lightweight CNN network that features a large effective receptive field (ERF) and demonstrates comparable or even better performance than Transformers while bearing less computational costs. Our key design is an efficient CNN block dubbed LaKD, equipped with a large kernel depth-wise convolution and spatial-channel mixing structure, attaining comparable or larger ERF than Transformers but with a smaller parameter scale. Specifically, we achieve +0.17dB / +0.43dB PSNR over the state-of-the-art Restormer on defocus / motion deblurring benchmark datasets with 32% fewer parameters and 39% fewer MACs. Extensive experiments demonstrate the superior performance of our network and the effectiveness of each module. Furthermore, we propose a compact and intuitive ERFMeter metric that quantitatively characterizes ERF, and shows a high correlation to the network performance. We hope this work can inspire the research community to further explore the pros and cons of CNN and Transformer architectures beyond image deblurring tasks.
comment: 30 pages (12 pages for the main manuscript and 18 for the supplementary materials)
☆ A Minimax Approach Against Multi-Armed Adversarial Attacks Detection
Multi-armed adversarial attacks, in which multiple algorithms and objective loss functions are simultaneously used at evaluation time, have been shown to be highly successful in fooling state-of-the-art adversarial examples detectors while requiring no specific side information about the detection mechanism. By formalizing the problem at hand, we can propose a solution that aggregates the soft-probability outputs of multiple pre-trained detectors according to a minimax approach. The proposed framework is mathematically sound, easy to implement, and modular, allowing for integrating existing or future detectors. Through extensive evaluation on popular datasets (e.g., CIFAR10 and SVHN), we show that our aggregation consistently outperforms individual state-of-the-art detectors against multi-armed adversarial attacks, making it an effective solution to improve the resilience of available methods.
comment: 10 pages, 13 figures, 14 tables
☆ Variational multichannel multiclass segmentation\endgraf using unsupervised lifting with CNNs
We propose an unsupervised image segmentation approach, that combines a variational energy functional and deep convolutional neural networks. The variational part is based on a recent multichannel multiphase Chan-Vese model, which is capable to extract useful information from multiple input images simultaneously. We implement a flexible multiclass segmentation method that divides a given image into $K$ different regions. We use convolutional neural networks (CNNs) targeting a pre-decomposition of the image. By subsequently minimising the segmentation functional, the final segmentation is obtained in a fully unsupervised manner. Special emphasis is given to the extraction of informative feature maps serving as a starting point for the segmentation. The initial results indicate that the proposed method is able to decompose and segment the different regions of various types of images, such as texture and medical images and compare its performance with another multiphase segmentation method.
comment: 20th INTERNATIONAL CONFERENCE OF NUMERICAL ANALYSIS AND APPLIED MATHEMATICS
☆ CosPGD: a unified white-box adversarial attack for pixel-wise prediction tasks
While neural networks allow highly accurate predictions in many tasks, their lack in robustness towards even slight input perturbations hampers their deployment in many real-world applications. Recent research towards evaluating the robustness of neural networks such as the seminal \emph{projected gradient descent} (PGD) attack and subsequent works and benchmarks have therefore drawn significant attention. Yet, such methods focus predominantly on classification tasks, while only a few approaches specifically address the analysis of pixel-wise prediction tasks such as semantic segmentation, optical flow, or disparity estimation. One notable exception is the recently proposed SegPGD attack, which could showcase the importance of pixel-wise attacks for evaluating semantic segmentation. While SegPGD is limited to pixel-wise classification (i.e. segmentation), in this work, we propose CosPGD, a novel white-box adversarial attack that allows to optimize dedicated attacks for any pixel-wise prediction task in a unified setting. It leverages the cosine similarity between the predictions and ground truth to extend directly from classification tasks to regression settings. Further, we empirically show the superior performance of CosPGD for semantic segmentation as well as for optical flow and disparity estimation.
☆ Oscillation-free Quantization for Low-bit Vision Transformers
Weight oscillation is an undesirable side effect of quantization-aware training, in which quantized weights frequently jump between two quantized levels, resulting in training instability and a sub-optimal final model. We discover that the learnable scaling factor, a widely-used $\textit{de facto}$ setting in quantization aggravates weight oscillation. In this study, we investigate the connection between the learnable scaling factor and quantized weight oscillation and use ViT as a case driver to illustrate the findings and remedies. In addition, we also found that the interdependence between quantized weights in $\textit{query}$ and $\textit{key}$ of a self-attention layer makes ViT vulnerable to oscillation. We, therefore, propose three techniques accordingly: statistical weight quantization ($\rm StatsQ$) to improve quantization robustness compared to the prevalent learnable-scale-based method; confidence-guided annealing ($\rm CGA$) that freezes the weights with $\textit{high confidence}$ and calms the oscillating weights; and $\textit{query}$-$\textit{key}$ reparameterization ($\rm QKR$) to resolve the query-key intertwined oscillation and mitigate the resulting gradient misestimation. Extensive experiments demonstrate that these proposed techniques successfully abate weight oscillation and consistently achieve substantial accuracy improvement on ImageNet. Specifically, our 2-bit DeiT-T/DeiT-S algorithms outperform the previous state-of-the-art by 9.8% and 7.7%, respectively. The code is included in the supplementary material and will be released.
comment: 11 pages, 11 figures
☆ Laplacian ICP for Progressive Registration of 3D Human Head Meshes
We present a progressive 3D registration framework that is a highly-efficient variant of classical non-rigid Iterative Closest Points (N-ICP). Since it uses the Laplace-Beltrami operator for deformation regularisation, we view the overall process as Laplacian ICP (L-ICP). This exploits a `small deformation per iteration' assumption and is progressively coarse-to-fine, employing an increasingly flexible deformation model, an increasing number of correspondence sets, and increasingly sophisticated correspondence estimation. Correspondence matching is only permitted within predefined vertex subsets derived from domain-specific feature extractors. Additionally, we present a new benchmark and a pair of evaluation metrics for 3D non-rigid registration, based on annotation transfer. We use this to evaluate our framework on a publicly-available dataset of 3D human head scans (Headspace). The method is robust and only requires a small fraction of the computation time compared to the most popular classical approach, yet has comparable registration performance.
comment: 7 pages, 6 figures
☆ Real-Time Image Demoireing on Mobile Devices ICLR 2023
Moire patterns appear frequently when taking photos of digital screens, drastically degrading the image quality. Despite the advance of CNNs in image demoireing, existing networks are with heavy design, causing redundant computation burden for mobile devices. In this paper, we launch the first study on accelerating demoireing networks and propose a dynamic demoireing acceleration method (DDA) towards a real-time deployment on mobile devices. Our stimulus stems from a simple-yet-universal fact that moire patterns often unbalancedly distribute across an image. Consequently, excessive computation is wasted upon non-moire areas. Therefore, we reallocate computation costs in proportion to the complexity of image patches. In order to achieve this aim, we measure the complexity of an image patch by designing a novel moire prior that considers both colorfulness and frequency information of moire patterns. Then, we restore image patches with higher-complexity using larger networks and the ones with lower-complexity are assigned with smaller networks to relieve the computation burden. At last, we train all networks in a parameter-shared supernet paradigm to avoid additional parameter burden. Extensive experiments on several benchmarks demonstrate the efficacy of our proposed DDA. In addition, the acceleration evaluated on the VIVO X80 Pro smartphone equipped with a chip of Snapdragon 8 Gen 1 shows that our method can drastically reduce the inference time, leading to a real-time image demoireing on mobile devices. Source codes and models are released at https://github.com/zyxxmu/DDA
comment: To appear in the eleventh International Conference on Learning Representations (ICLR 2023)
☆ Model Stitching and Visualization How GAN Generators can Invert Networks in Real-Time
Critical applications, such as in the medical field, require the rapid provision of additional information to interpret decisions made by deep learning methods. In this work, we propose a fast and accurate method to visualize activations of classification and semantic segmentation networks by stitching them with a GAN generator utilizing convolutions. We test our approach on images of animals from the AFHQ wild dataset and real-world digital pathology scans of stained tissue samples. Our method provides comparable results to established gradient descent methods on these datasets while running about two orders of magnitude faster.
☆ Guaranteed Tensor Recovery Fused Low-rankness and Smoothness
The tensor data recovery task has thus attracted much research attention in recent years. Solving such an ill-posed problem generally requires to explore intrinsic prior structures underlying tensor data, and formulate them as certain forms of regularization terms for guiding a sound estimate of the restored tensor. Recent research have made significant progress by adopting two insightful tensor priors, i.e., global low-rankness (L) and local smoothness (S) across different tensor modes, which are always encoded as a sum of two separate regularization terms into the recovery models. However, unlike the primary theoretical developments on low-rank tensor recovery, these joint L+S models have no theoretical exact-recovery guarantees yet, making the methods lack reliability in real practice. To this crucial issue, in this work, we build a unique regularization term, which essentially encodes both L and S priors of a tensor simultaneously. Especially, by equipping this single regularizer into the recovery models, we can rigorously prove the exact recovery guarantees for two typical tensor recovery tasks, i.e., tensor completion (TC) and tensor robust principal component analysis (TRPCA). To the best of our knowledge, this should be the first exact-recovery results among all related L+S methods for tensor recovery. Significant recovery accuracy improvements over many other SOTA methods in several TC and TRPCA tasks with various kinds of visual tensor data are observed in extensive experiments. Typically, our method achieves a workable performance when the missing rate is extremely large, e.g., 99.5%, for the color image inpainting task, while all its peers totally fail in such challenging case.
☆ This Intestine Does Not Exist: Multiscale Residual Variational Autoencoder for Realistic Wireless Capsule Endoscopy Image Generation
Medical image synthesis has emerged as a promising solution to address the limited availability of annotated medical data needed for training machine learning algorithms in the context of image-based Clinical Decision Support (CDS) systems. To this end, Generative Adversarial Networks (GANs) have been mainly applied to support the algorithm training process by generating synthetic images for data augmentation. However, in the field of Wireless Capsule Endoscopy (WCE), the limited content diversity and size of existing publicly available annotated datasets, adversely affect both the training stability and synthesis performance of GANs. Aiming to a viable solution for WCE image synthesis, a novel Variational Autoencoder architecture is proposed, namely "This Intestine Does not Exist" (TIDE). The proposed architecture comprises multiscale feature extraction convolutional blocks and residual connections, which enable the generation of high-quality and diverse datasets even with a limited number of training images. Contrary to the current approaches, which are oriented towards the augmentation of the available datasets, this study demonstrates that using TIDE, real WCE datasets can be fully substituted by artificially generated ones, without compromising classification performance. Furthermore, qualitative and user evaluation studies by experienced WCE specialists, validate from a medical viewpoint that both the normal and abnormal WCE images synthesized by TIDE are sufficiently realistic.
comment: 10 pages
☆ LipFormer: Learning to Lipread Unseen Speakers based on Visual-Landmark Transformers
Lipreading refers to understanding and further translating the speech of a speaker in the video into natural language. State-of-the-art lipreading methods excel in interpreting overlap speakers, i.e., speakers appear in both training and inference sets. However, generalizing these methods to unseen speakers incurs catastrophic performance degradation due to the limited number of speakers in training bank and the evident visual variations caused by the shape/color of lips for different speakers. Therefore, merely depending on the visible changes of lips tends to cause model overfitting. To address this problem, we propose to use multi-modal features across visual and landmarks, which can describe the lip motion irrespective to the speaker identities. Then, we develop a sentence-level lipreading framework based on visual-landmark transformers, namely LipFormer. Specifically, LipFormer consists of a lip motion stream, a facial landmark stream, and a cross-modal fusion. The embeddings from the two streams are produced by self-attention, which are fed to the cross-attention module to achieve the alignment between visuals and landmarks. Finally, the resulting fused features can be decoded to output texts by a cascade seq2seq model. Experiments demonstrate that our method can effectively enhance the model generalization to unseen speakers.
comment: Under review
☆ Efficient End-to-End Video Question Answering with Pyramidal Multimodal Transformer AAAI 2023
This paper presents a new method for end-to-end Video Question Answering (VideoQA), aside from the current popularity of using large-scale pre-training with huge feature extractors. We achieve this with a pyramidal multimodal transformer (PMT) model, which simply incorporates a learnable word embedding layer, a few convolutional and transformer layers. We use the anisotropic pyramid to fulfill video-language interactions across different spatio-temporal scales. In addition to the canonical pyramid, which includes both bottom-up and top-down pathways with lateral connections, novel strategies are proposed to decompose the visual feature stream into spatial and temporal sub-streams at different scales and implement their interactions with the linguistic semantics while preserving the integrity of local and global semantics. We demonstrate better or on-par performances with high computational efficiency against state-of-the-art methods on five VideoQA benchmarks. Our ablation study shows the scalability of our model that achieves competitive results for text-to-video retrieval by leveraging feature extractors with reusable pre-trained weights, and also the effectiveness of the pyramid.
comment: Accepted by AAAI 2023
☆ Weakly-Supervised 3D Medical Image Segmentation using Geometric Prior and Contrastive Similarity
Medical image segmentation is almost the most important pre-processing procedure in computer-aided diagnosis but is also a very challenging task due to the complex shapes of segments and various artifacts caused by medical imaging, (i.e., low-contrast tissues, and non-homogenous textures). In this paper, we propose a simple yet effective segmentation framework that incorporates the geometric prior and contrastive similarity into the weakly-supervised segmentation framework in a loss-based fashion. The proposed geometric prior built on point cloud provides meticulous geometry to the weakly-supervised segmentation proposal, which serves as better supervision than the inherent property of the bounding-box annotation (i.e., height and width). Furthermore, we propose contrastive similarity to encourage organ pixels to gather around in the contrastive embedding space, which helps better distinguish low-contrast tissues. The proposed contrastive embedding space can make up for the poor representation of the conventionally-used gray space. Extensive experiments are conducted to verify the effectiveness and the robustness of the proposed weakly-supervised segmentation framework. The proposed framework is superior to state-of-the-art weakly-supervised methods on the following publicly accessible datasets: LiTS 2017 Challenge, KiTS 2021 Challenge, and LPBA40. We also dissect our method and evaluate the performance of each component.
comment: Weakly-supervised Segmentation, Medical Image Segmentation, Contrastive Similarity, Geometric Prior, Point Cloud
☆ Transform, Contrast and Tell: Coherent Entity-Aware Multi-Image Captioning
Coherent entity-aware multi-image captioning aims to generate coherent captions for multiple adjacent images in a news document. There are coherence relationships among adjacent images because they often describe same entities or events. These relationships are important for entity-aware multi-image captioning, but are neglected in entity-aware single-image captioning. Most existing work focuses on single-image captioning, while multi-image captioning has not been explored before. Hence, this paper proposes a coherent entity-aware multi-image captioning model by making use of coherence relationships. The model consists of a Transformer-based caption generation model and two types of contrastive learning-based coherence mechanisms. The generation model generates the caption by paying attention to the image and the accompanying text. The horizontal coherence mechanism aims to the make the caption coherent with captions of adjacent images. The vertical coherence mechanism aims to make the caption coherent with the image and the accompanying text. To evaluate coherence between captions, two coherence evaluation metrics are proposed. The new dataset DM800K is constructed that has more images per document than two existing datasets GoodNews and NYT800K, and are more suitable for multi-image captioning. Experiments on three datasets show the proposed captioning model outperforms 6 baselines according to single-image captioning evaluations, and the generated captions are more coherent than that of baselines according to coherence evaluations and human evaluations.
comment: 28 pages, 9 tables, 3 figures
☆ Learning to Agree on Vision Attention for Visual Commonsense Reasoning
Visual Commonsense Reasoning (VCR) remains a significant yet challenging research problem in the realm of visual reasoning. A VCR model generally aims at answering a textual question regarding an image, followed by the rationale prediction for the preceding answering process. Though these two processes are sequential and intertwined, existing methods always consider them as two independent matching-based instances. They, therefore, ignore the pivotal relationship between the two processes, leading to sub-optimal model performance. This paper presents a novel visual attention alignment method to efficaciously handle these two processes in a unified framework. To achieve this, we first design a re-attention module for aggregating the vision attention map produced in each process. Thereafter, the resultant two sets of attention maps are carefully aligned to guide the two processes to make decisions based on the same image regions. We apply this method to both conventional attention and the recent Transformer models and carry out extensive experiments on the VCR benchmark dataset. The results demonstrate that with the attention alignment module, our method achieves a considerable improvement over the baseline methods, evidently revealing the feasibility of the coupling of the two processes as well as the effectiveness of the proposed method.
☆ Knowledge Distillation in Vision Transformers: A Critical Review
In Natural Language Processing (NLP), Transformers have already revolutionized the field by utilizing an attention-based encoder-decoder model. Recently, some pioneering works have employed Transformer-like architectures in Computer Vision (CV) and they have reported outstanding performance of these architectures in tasks such as image classification, object detection, and semantic segmentation. Vision Transformers (ViTs) have demonstrated impressive performance improvements over Convolutional Neural Networks (CNNs) due to their competitive modelling capabilities. However, these architectures demand massive computational resources which makes these models difficult to be deployed in the resource-constrained applications. Many solutions have been developed to combat this issue, such as compressive transformers and compression functions such as dilated convolution, min-max pooling, 1D convolution, etc. Model compression has recently attracted considerable research attention as a potential remedy. A number of model compression methods have been proposed in the literature such as weight quantization, weight multiplexing, pruning and Knowledge Distillation (KD). However, techniques like weight quantization, pruning and weight multiplexing typically involve complex pipelines for performing the compression. KD has been found to be a simple and much effective model compression technique that allows a relatively simple model to perform tasks almost as accurately as a complex model. This paper discusses various approaches based upon KD for effective compression of ViT models. The paper elucidates the role played by KD in reducing the computational and memory requirements of these models. The paper also presents the various challenges faced by ViTs that are yet to be resolved.
comment: 28pages, 16 figures
♻ ☆ Continual Segment: Towards a Single, Unified and Accessible Continual Segmentation Model of 143 Whole-body Organs in CT Scans
Deep learning empowers the mainstream medical image segmentation methods. Nevertheless current deep segmentation approaches are not capable of efficiently and effectively adapting and updating the trained models when new incremental segmentation classes (along with new training datasets or not) are required to be added. In real clinical environment, it can be preferred that segmentation models could be dynamically extended to segment new organs/tumors without the (re-)access to previous training datasets due to obstacles of patient privacy and data storage. This process can be viewed as a continual semantic segmentation (CSS) problem, being understudied for multi-organ segmentation. In this work, we propose a new architectural CSS learning framework to learn a single deep segmentation model for segmenting a total of 143 whole-body organs. Using the encoder/decoder network structure, we demonstrate that a continually-trained then frozen encoder coupled with incrementally-added decoders can extract and preserve sufficiently representative image features for new classes to be subsequently and validly segmented. To maintain a single network model complexity, we trim each decoder progressively using neural architecture search and teacher-student based knowledge distillation. To incorporate with both healthy and pathological organs appearing in different datasets, a novel anomaly-aware and confidence learning module is proposed to merge the overlapped organ predictions, originated from different decoders. Trained and validated on 3D CT scans of 2500+ patients from four datasets, our single network can segment total 143 whole-body organs with very high accuracy, closely reaching the upper bound performance level by training four separate segmentation models (i.e., one model per dataset/task).
♻ ☆ Detecting Disengagement in Virtual Learning as an Anomaly using Temporal Convolutional Network Autoencoder
Student engagement is an important factor in meeting the goals of virtual learning programs. Automatic measurement of student engagement provides helpful information for instructors to meet learning program objectives and individualize program delivery. Many existing approaches solve video-based engagement measurement using the traditional frameworks of binary classification (classifying video snippets into engaged or disengaged classes), multi-class classification (classifying video snippets into multiple classes corresponding to different levels of engagement), or regression (estimating a continuous value corresponding to the level of engagement). However, we observe that while the engagement behaviour is mostly well-defined (e.g., focused, not distracted), disengagement can be expressed in various ways. In addition, in some cases, the data for disengaged classes may not be sufficient to train generalizable binary or multi-class classifiers. To handle this situation, in this paper, for the first time, we formulate detecting disengagement in virtual learning as an anomaly detection problem. We design various autoencoders, including temporal convolutional network autoencoder, long-short-term memory autoencoder, and feedforward autoencoder using different behavioral and affect features for video-based student disengagement detection. The result of our experiments on two publicly available student engagement datasets, DAiSEE and EmotiW, shows the superiority of the proposed approach for disengagement detection as an anomaly compared to binary classifiers for classifying videos into engaged versus disengaged classes (with an average improvement of 9% on the area under the curve of the receiver operating characteristic curve and 22% on the area under the curve of the precision-recall curve).
♻ ☆ Towards Improving the Generation Quality of Autoregressive Slot VAEs
Unconditional scene inference and generation are challenging to learn jointly with a single compositional model. Despite encouraging progress on models that extract object-centric representations ("slots") from images, unconditional generation of scenes from slots has received less attention. This is primarily because learning the multi-object relations necessary to imagine coherent scenes is difficult. We hypothesize that most existing slot-based models have a limited ability to learn object correlations. We propose two improvements that strengthen slot correlation learning. The first is to condition the slots on a global, scene-level variable that captures higher-order correlations between slots. Second, we address the fundamental lack of a canonical order for objects by proposing to learn a consistent order to use for the autoregressive generation of scene objects. Specifically, we train an autoregressive slot prior to sequentially generate scene objects following the learned order. Slot inference entails estimating a randomly ordered set of slots using existing approaches for extracting slots from images, then aligning those slots to ordered slots generated autoregressively with the prior. Our experiments across three multi-object environments demonstrate clear gains in scene generation quality. Detailed ablation studies are also provided that validate the two proposed improvements.
comment: 39 pages, 16 figures. Under review. Code and videos available at https://github.com/pemami4911/segregate-relate-imagine
♻ ☆ Large-scale Building Damage Assessment using a Novel Hierarchical Transformer Architecture on Satellite Images
This paper presents \dahitra, a novel deep-learning model with hierarchical transformers to classify building damages based on satellite images in the aftermath of natural disasters. Satellite imagery provides real-time and high-coverage information and offers opportunities to inform large-scale post-disaster building damage assessment, which is critical for rapid emergency response. In this work, a novel transformer-based network is proposed for assessing building damage. This network leverages hierarchical spatial features of multiple resolutions and captures the temporal differences in the feature domain after applying a transformer encoder on the spatial features. The proposed network achieves state-of-the-art performance when tested on a large-scale disaster damage dataset (xBD) for building localization and damage classification, as well as on LEVIR-CD dataset for change detection tasks. In addition, this work introduces a new high-resolution satellite imagery dataset, Ida-BD (related to 2021 Hurricane Ida in Louisiana in 2021) for domain adaptation. Further, it demonstrates an approach of using this dataset by adapting the model with limited fine-tuning and hence applying the model to newly damaged areas with scarce data.
♻ ☆ GEDI: GEnerative and DIscriminative Training for Self-Supervised Learning
Self-supervised learning is a popular and powerful method for utilizing large amounts of unlabeled data, for which a wide variety of training objectives have been proposed in the literature. In this study, we perform a Bayesian analysis of state-of-the-art self-supervised learning objectives and propose a unified formulation based on likelihood learning. Our analysis suggests a simple method for integrating self-supervised learning with generative models, allowing for the joint training of these two seemingly distinct approaches. We refer to this combined framework as GEDI, which stands for GEnerative and DIscriminative training. Additionally, we demonstrate an instantiation of the GEDI framework by integrating an energy-based model with a cluster-based self-supervised learning model. Through experiments on synthetic and real-world data, including SVHN, CIFAR10, and CIFAR100, we show that GEDI outperforms existing self-supervised learning strategies in terms of clustering performance by a wide margin. We also demonstrate that GEDI can be integrated into a neural-symbolic framework to address tasks in the small data regime, where it can use logical constraints to further improve clustering and classification performance.
comment: Fixed typos/cleaned the experimental section
♻ ☆ Trackerless freehand ultrasound with sequence modelling and auxiliary transformation over past and future frames
Three-dimensional (3D) freehand ultrasound (US) reconstruction without a tracker can be advantageous over its two-dimensional or tracked counterparts in many clinical applications. In this paper, we propose to estimate 3D spatial transformation between US frames from both past and future 2D images, using feed-forward and recurrent neural networks (RNNs). With the temporally available frames, a further multi-task learning algorithm is proposed to utilise a large number of auxiliary transformation-predicting tasks between them. Using more than 40,000 US frames acquired from 228 scans on 38 forearms of 19 volunteers in a volunteer study, the hold-out test performance is quantified by frame prediction accuracy, volume reconstruction overlap, accumulated tracking error and final drift, based on ground-truth from an optical tracker. The results show the importance of modelling the temporal-spatially correlated input frames as well as output transformations, with further improvement owing to additional past and/or future frames. The best performing model was associated with predicting transformation between moderately-spaced frames, with an interval of less than ten frames at 20 frames per second (fps). Little benefit was observed by adding frames more than one second away from the predicted transformation, with or without LSTM-based RNNs. Interestingly, with the proposed approach, explicit within-sequence loss that encourages consistency in composing transformations or minimises accumulated error may no longer be required. The implementation code and volunteer data will be made publicly available ensuring reproducibility and further research.
comment: Accepted to IEEE International Symposium on Biomedical Imaging (ISBI) 2023
♻ ☆ Bayesian Metric Learning for Uncertainty Quantification in Image Retrieval
We propose the first Bayesian encoder for metric learning. Rather than relying on neural amortization as done in prior works, we learn a distribution over the network weights with the Laplace Approximation. We actualize this by first proving that the contrastive loss is a valid log-posterior. We then propose three methods that ensure a positive definite Hessian. Lastly, we present a novel decomposition of the Generalized Gauss-Newton approximation. Empirically, we show that our Laplacian Metric Learner (LAM) estimates well-calibrated uncertainties, reliably detects out-of-distribution examples, and yields state-of-the-art predictive performance.
comment: Code: https://github.com/FrederikWarburg/bayesian-metric-learning
♻ ☆ Progressive Training of A Two-Stage Framework for Video Restoration CVPR
As a widely studied task, video restoration aims to enhance the quality of the videos with multiple potential degradations, such as noises, blurs and compression artifacts. Among video restorations, compressed video quality enhancement and video super-resolution are two of the main tacks with significant values in practical scenarios. Recently, recurrent neural networks and transformers attract increasing research interests in this field, due to their impressive capability in sequence-to-sequence modeling. However, the training of these models is not only costly but also relatively hard to converge, with gradient exploding and vanishing problems. To cope with these problems, we proposed a two-stage framework including a multi-frame recurrent network and a single-frame transformer. Besides, multiple training strategies, such as transfer learning and progressive training, are developed to shorten the training time and improve the model performance. Benefiting from the above technical contributions, our solution wins two champions and a runner-up in the NTIRE 2022 super-resolution and quality enhancement of compressed video challenges. Code is available at https://github.com/ryanxingql/winner-ntire22-vqe.
comment: Winning two championships and one runner-up in the NTIRE 2022 challenge on super-resolution and quality enhancement of compressed video; Accepted to CVPRW 2022
♻ ☆ ImaginE: An Imagination-Based Automatic Evaluation Metric for Natural Language Generation EACL 2023
Automatic evaluations for natural language generation (NLG) conventionally rely on token-level or embedding-level comparisons with text references. This differs from human language processing, for which visual imagination often improves comprehension. In this work, we propose ImaginE, an imagination-based automatic evaluation metric for natural language generation. With the help of StableDiffusion, a state-of-the-art text-to-image generator, we automatically generate an image as the embodied imagination for the text snippet and compute the imagination similarity using contextual embeddings. Experiments spanning several text generation tasks demonstrate that adding machine-generated images with our ImaginE displays great potential in introducing multi-modal information into NLG evaluation, and improves existing automatic metrics' correlations with human similarity judgments in both reference-based and reference-free evaluation scenarios.
comment: EACL 2023
♻ ☆ Distilling Cognitive Backdoor Patterns within an Image ICLR2023
This paper proposes a simple method to distill and detect backdoor patterns within an image: \emph{Cognitive Distillation} (CD). The idea is to extract the "minimal essence" from an input image responsible for the model's prediction. CD optimizes an input mask to extract a small pattern from the input image that can lead to the same model output (i.e., logits or deep features). The extracted pattern can help understand the cognitive mechanism of a model on clean vs. backdoor images and is thus called a \emph{Cognitive Pattern} (CP). Using CD and the distilled CPs, we uncover an interesting phenomenon of backdoor attacks: despite the various forms and sizes of trigger patterns used by different attacks, the CPs of backdoor samples are all surprisingly and suspiciously small. One thus can leverage the learned mask to detect and remove backdoor examples from poisoned training datasets. We conduct extensive experiments to show that CD can robustly detect a wide range of advanced backdoor attacks. We also show that CD can potentially be applied to help detect potential biases from face datasets. Code is available at \url{https://github.com/HanxunH/CognitiveDistillation}.
comment: ICLR2023
♻ ☆ Comparison of modern open-source visual SLAM approaches
SLAM is one of the most fundamental areas of research in robotics and computer vision. State of the art solutions has advanced significantly in terms of accuracy and stability. Unfortunately, not all the approaches are available as open-source solutions and free to use. The results of some of them are difficult to reproduce, and there is a lack of comparison on common datasets. In our work, we make a comparative analysis of state of the art open-source methods. We assess the algorithms based on accuracy, computational performance, robustness, and fault tolerance. Moreover, we present a comparison of datasets as well as an analysis of algorithms from a practical point of view. The findings of the work raise several crucial questions for SLAM researchers.
comment: Preprint, 19 pages
Information Retrieval 5
☆ Feature Representation Learning for Click-through Rate Prediction: A Review and New Perspectives IJCAI 2023
Representation learning has been a critical topic in machine learning. In Click-through Rate Prediction, most features are represented as embedding vectors and learned simultaneously with other parameters in the model. With the development of CTR models, feature representation learning has become a trending topic and has been extensively studied by both industrial and academic researchers in recent years. This survey aims at summarizing the feature representation learning in a broader picture and pave the way for future research. To achieve such a goal, we first present a taxonomy of current research methods on feature representation learning following two main issues: (i) which feature to represent and (ii) how to represent these features. Then we give a detailed description of each method regarding these two issues. Finally, the review concludes with a discussion on the future directions of this field.
comment: Submitted to IJCAI 2023 Survey Track
☆ Contrastive Collaborative Filtering for Cold-Start Item Recommendation WWW '23
The cold-start problem is a long-standing challenge in recommender systems. As a promising solution, content-based generative models usually project a cold-start item's content onto a warm-start item embedding to capture collaborative signals from item content so that collaborative filtering can be applied. However, since the training of the cold-start recommendation models is conducted on warm datasets, the existent methods face the issue that the collaborative embeddings of items will be blurred, which significantly degenerates the performance of cold-start item recommendation. To address this issue, we propose a novel model called Contrastive Collaborative Filtering for Cold-start item Recommendation (CCFCRec), which capitalizes on the co-occurrence collaborative signals in warm training data to alleviate the issue of blurry collaborative embeddings for cold-start item recommendation. In particular, we devise a contrastive collaborative filtering (CF) framework, consisting of a content CF module and a co-occurrence CF module to generate the content-based collaborative embedding and the co-occurrence collaborative embedding for a training item, respectively. During the joint training of the two CF modules, we apply a contrastive learning between the two collaborative embeddings, by which the knowledge about the co-occurrence signals can be indirectly transferred to the content CF module, so that the blurry collaborative embeddings can be rectified implicitly by the memorized co-occurrence collaborative signals during the applying phase. Together with the sound theoretical analysis, the extensive experiments conducted on real datasets demonstrate the superiority of the proposed model. The codes and datasets are available on https://github.com/zzhin/CCFCRec.
comment: This paper has been accepted by WWW '23
☆ Personalized Graph Signal Processing for Collaborative Filtering WWW 2023
The collaborative filtering (CF) problem with only user-item interaction information can be solved by graph signal processing (GSP), which uses low-pass filters to smooth the observed interaction signals on the similarity graph to obtain the prediction signals. However, the interaction signal may not be sufficient to accurately characterize user interests and the low-pass filters may ignore the useful information contained in the high-frequency component of the observed signals, resulting in suboptimal accuracy. To this end, we propose a personalized graph signal processing (PGSP) method for collaborative filtering. Firstly, we design the personalized graph signal containing richer user information and construct an augmented similarity graph containing more graph topology information, to more effectively characterize user interests. Secondly, we devise a mixed-frequency graph filter to introduce useful information in the high-frequency components of the observed signals by combining an ideal low-pass filter that smooths signals globally and a linear low-pass filter that smooths signals locally. Finally, we combine the personalized graph signal, the augmented similarity graph and the mixed-frequency graph filter by proposing a pipeline consisting of three key steps: pre-processing, graph convolution and post-processing. Extensive experiments show that PGSP can achieve superior accuracy compared with state-of-the-art CF methods and, as a nonparametric method, PGSP has very high training efficiency.
comment: Accepted by WWW 2023, 9 pages
♻ ☆ CitationSum: Citation-aware Graph Contrastive Learning for Scientific Paper Summarization WWW2023
Citation graphs can be helpful in generating high-quality summaries of scientific papers, where references of a scientific paper and their correlations can provide additional knowledge for contextualising its background and main contributions. Despite the promising contributions of citation graphs, it is still challenging to incorporate them into summarization tasks. This is due to the difficulty of accurately identifying and leveraging relevant content in references for a source paper, as well as capturing their correlations of different intensities. Existing methods either ignore references or utilize only abstracts indiscriminately from them, failing to tackle the challenge mentioned above. To fill that gap, we propose a novel citation-aware scientific paper summarization framework based on citation graphs, able to accurately locate and incorporate the salient contents from references, as well as capture varying relevance between source papers and their references. Specifically, we first build a domain-specific dataset PubMedCite with about 192K biomedical scientific papers and a large citation graph preserving 917K citation relationships between them. It is characterized by preserving the salient contents extracted from full texts of references, and the weighted correlation between the salient contents of references and the source paper. Based on it, we design a self-supervised citation-aware summarization framework (CitationSum) with graph contrastive learning, which boosts the summarization generation by efficiently fusing the salient information in references with source paper contents under the guidance of their correlations. Experimental results show that our model outperforms the state-of-the-art methods, due to efficiently leveraging the information of references and citation correlations.
comment: accepted to WWW2023
♻ ☆ A Concept Knowledge Graph for User Next Intent Prediction at Alipay
This paper illustrates the technologies of user next intent prediction with a concept knowledge graph. The system has been deployed on the Web at Alipay, serving more than 100 million daily active users. To explicitly characterize user intent, we propose \textbf{AlipayKG}, which is an offline concept knowledge graph in the Life-Service domain modeling the historical behaviors of users, the rich content interacted by users and the relations between them. We further introduce a Transformer-based model which integrates expert rules from the knowledge graph to infer the online user's next intent. Experimental results demonstrate that the proposed system can effectively enhance the performance of the downstream tasks while retaining explainability.
comment: Work in progress
Multimedia 3
☆ An analysis of the technology acceptance model in understanding university students behavioral intention to use metaverse technologies
Metaverse can be applied in several aspects of life such as the Economy, finance, social life, working environment, healthcare, real estate, and education. In the last 2 and a half years, during the COVID-19 pandemic, universities made immediate use of learning technologies, providing students with access to online learning content and platforms. Previous considerations on how to better integrate technology into universities or how the institutions can be better prepared in terms of infrastructure vanished almost immediately due to the necessity of immediate actions towards the need for social distance and global health. The present study proposes a framework for university students metaverse technologies in education acceptance and intention to use. The present study develops a structural model of MetaEducation acceptance. This model will be useful to university managers, policymakers, and professors to better incorporate the upcoming metaverse technology. The present study tests (if supported) the correlations among the aforementioned constructs. Preliminary results show hesitance to use MetaEducation technologies from university students. Self-efficacy and Subjective Norms affect Attitude and Perceived Usefulness positively, but on the other side, there is no strong correlation between Perceived Ease of Use and Attitude or Perceived Usefulness and Attitude. Authors believe that the weak ties among the study constructs have to do with the lack of knowledge of what really MetaEducation really is, and which are its advantages of use.
comment: 4 pages, 3 figures, conference article
☆ LipFormer: Learning to Lipread Unseen Speakers based on Visual-Landmark Transformers
Lipreading refers to understanding and further translating the speech of a speaker in the video into natural language. State-of-the-art lipreading methods excel in interpreting overlap speakers, i.e., speakers appear in both training and inference sets. However, generalizing these methods to unseen speakers incurs catastrophic performance degradation due to the limited number of speakers in training bank and the evident visual variations caused by the shape/color of lips for different speakers. Therefore, merely depending on the visible changes of lips tends to cause model overfitting. To address this problem, we propose to use multi-modal features across visual and landmarks, which can describe the lip motion irrespective to the speaker identities. Then, we develop a sentence-level lipreading framework based on visual-landmark transformers, namely LipFormer. Specifically, LipFormer consists of a lip motion stream, a facial landmark stream, and a cross-modal fusion. The embeddings from the two streams are produced by self-attention, which are fed to the cross-attention module to achieve the alignment between visuals and landmarks. Finally, the resulting fused features can be decoded to output texts by a cascade seq2seq model. Experiments demonstrate that our method can effectively enhance the model generalization to unseen speakers.
comment: Under review
♻ ☆ Progressive Training of A Two-Stage Framework for Video Restoration CVPR
As a widely studied task, video restoration aims to enhance the quality of the videos with multiple potential degradations, such as noises, blurs and compression artifacts. Among video restorations, compressed video quality enhancement and video super-resolution are two of the main tacks with significant values in practical scenarios. Recently, recurrent neural networks and transformers attract increasing research interests in this field, due to their impressive capability in sequence-to-sequence modeling. However, the training of these models is not only costly but also relatively hard to converge, with gradient exploding and vanishing problems. To cope with these problems, we proposed a two-stage framework including a multi-frame recurrent network and a single-frame transformer. Besides, multiple training strategies, such as transfer learning and progressive training, are developed to shorten the training time and improve the model performance. Benefiting from the above technical contributions, our solution wins two champions and a runner-up in the NTIRE 2022 super-resolution and quality enhancement of compressed video challenges. Code is available at https://github.com/ryanxingql/winner-ntire22-vqe.
comment: Winning two championships and one runner-up in the NTIRE 2022 challenge on super-resolution and quality enhancement of compressed video; Accepted to CVPRW 2022
Computation and Language 47
☆ GLADIS: A General and Large Acronym Disambiguation Benchmark EACL 23
Acronym Disambiguation (AD) is crucial for natural language understanding on various sources, including biomedical reports, scientific papers, and search engine queries. However, existing acronym disambiguation benchmarks and tools are limited to specific domains, and the size of prior benchmarks is rather small. To accelerate the research on acronym disambiguation, we construct a new benchmark named GLADIS with three components: (1) a much larger acronym dictionary with 1.5M acronyms and 6.4M long forms; (2) a pre-training corpus with 160 million sentences; (3) three datasets that cover the general, scientific, and biomedical domains. We then pre-train a language model, \emph{AcroBERT}, on our constructed corpus for general acronym disambiguation, and show the challenges and values of our new benchmark.
comment: EACL 23
☆ Generalizing to Unseen Elements: A Survey on Knowledge Extrapolation for Knowledge Graphs
Knowledge graphs (KGs) have become effective knowledge resources in diverse applications, and knowledge graph embedding (KGE) methods have attracted increasing attention in recent years. However, it's still challenging for conventional KGE methods to handle unseen entities or relations during the model test. Much effort has been made in various fields of KGs to address this problem. In this paper, we use a set of general terminologies to unify these methods and refer to them as Knowledge Extrapolation. We comprehensively summarize these methods classified by our proposed taxonomy and describe their correlations. Next, we introduce the benchmarks and provide comparisons of these methods from aspects that are not reflected by the taxonomy. Finally, we suggest some potential directions for future research.
☆ Entity-Agnostic Representation Learning for Parameter-Efficient Knowledge Graph Embedding AAAI 2023
We propose an entity-agnostic representation learning method for handling the problem of inefficient parameter storage costs brought by embedding knowledge graphs. Conventional knowledge graph embedding methods map elements in a knowledge graph, including entities and relations, into continuous vector spaces by assigning them one or multiple specific embeddings (i.e., vector representations). Thus the number of embedding parameters increases linearly as the growth of knowledge graphs. In our proposed model, Entity-Agnostic Representation Learning (EARL), we only learn the embeddings for a small set of entities and refer to them as reserved entities. To obtain the embeddings for the full set of entities, we encode their distinguishable information from their connected relations, k-nearest reserved entities, and multi-hop neighbors. We learn universal and entity-agnostic encoders for transforming distinguishable information into entity embeddings. This approach allows our proposed EARL to have a static, efficient, and lower parameter count than conventional knowledge graph embedding methods. Experimental results show that EARL uses fewer parameters and performs better on link prediction tasks than baselines, reflecting its parameter efficiency.
comment: Accepted to AAAI 2023 conference
☆ A Case Study for Compliance as Code with Graphs and Language Models: Public release of the Regulatory Knowledge Graph
The paper presents a study on using language models to automate the construction of executable Knowledge Graph (KG) for compliance. The paper focuses on Abu Dhabi Global Market regulations and taxonomy, involves manual tagging a portion of the regulations, training BERT-based models, which are then applied to the rest of the corpus. Coreference resolution and syntax analysis were used to parse the relationships between the tagged entities and to form KG stored in a Neo4j database. The paper states that the use of machine learning models released by regulators to automate the interpretation of rules is a vital step towards compliance automation, demonstrates the concept querying with Cypher, and states that the produced sub-graphs combined with Graph Neural Networks (GNN) will achieve expandability in judgment automation systems. The graph is open sourced on GitHub to provide structured data for future advancements in the field.
comment: 7 pages on RegKG step 1 in collaboration with ADGM
☆ Investigating Stylistic Profiles for the Task of Empathy Classification in Medical Narrative Essays ACL
One important aspect of language is how speakers generate utterances and texts to convey their intended meanings. In this paper, we bring various aspects of the Construction Grammar (CxG) and the Systemic Functional Grammar (SFG) theories in a deep learning computational framework to model empathic language. Our corpus consists of 440 essays written by premed students as narrated simulated patient-doctor interactions. We start with baseline classifiers (state-of-the-art recurrent neural networks and transformer models). Then, we enrich these models with a set of linguistic constructions proving the importance of this novel approach to the task of empathy classification for this dataset. Our results indicate the potential of such constructions to contribute to the overall empathy profile of first-person narrative essays.
comment: 12 pages, 5 figures; This paper will appear in the ACL Anthology (Association for Computational Linguistics) and will be presented at the Construction Grammars and NLP (CxGs+NLP) Workshop, the Georgetown University Round Table (GURT), March 2023
☆ Lexical Simplification using multi level and modular approach
Text Simplification is an ongoing problem in Natural Language Processing, solution to which has varied implications. In conjunction with the TSAR-2022 Workshop @EMNLP2022 Lexical Simplification is the process of reducing the lexical complexity of a text by replacing difficult words with easier to read (or understand) expressions while preserving the original information and meaning. This paper explains the work done by our team "teamPN" for English sub task. We created a modular pipeline which combines modern day transformers based models with traditional NLP methods like paraphrasing and verb sense disambiguation. We created a multi level and modular pipeline where the target text is treated according to its semantics(Part of Speech Tag). Pipeline is multi level as we utilize multiple source models to find potential candidates for replacement, It is modular as we can switch the source models and their weight-age in the final re-ranking.
☆ Mitigating Data Scarcity for Large Language Models
In recent years, pretrained neural language models (PNLMs) have taken the field of natural language processing by storm, achieving new benchmarks and state-of-the-art performances. These models often rely heavily on annotated data, which may not always be available. Data scarcity are commonly found in specialized domains, such as medical, or in low-resource languages that are underexplored by AI research. In this dissertation, we focus on mitigating data scarcity using data augmentation and neural ensemble learning techniques for neural language models. In both research directions, we implement neural network algorithms and evaluate their impact on assisting neural language models in downstream NLP tasks. Specifically, for data augmentation, we explore two techniques: 1) creating positive training data by moving an answer span around its original context and 2) using text simplification techniques to introduce a variety of writing styles to the original training data. Our results indicate that these simple and effective solutions improve the performance of neural language models considerably in low-resource NLP domains and tasks. For neural ensemble learning, we use a multilabel neural classifier to select the best prediction outcome from a variety of individual pretrained neural language models trained for a low-resource medical text simplification task.
comment: 155 pages, 26 tables, 11 figures
☆ LIQUID: A Framework for List Question Answering Dataset Generation AAAI 2023
Question answering (QA) models often rely on large-scale training datasets, which necessitates the development of a data generation framework to reduce the cost of manual annotations. Although several recent studies have aimed to generate synthetic questions with single-span answers, no study has been conducted on the creation of list questions with multiple, non-contiguous spans as answers. To address this gap, we propose \ours, an automated framework for generating list QA datasets from unlabeled corpora. We first convert a passage from Wikipedia or PubMed into a summary and extract named entities from the summarized text as candidate answers. This allows us to select answers that are semantically correlated in context and is, therefore, suitable for constructing list questions. We then create questions using an off-the-shelf question generator with the extracted entities and original passage. Finally, iterative filtering and answer expansion are performed to ensure the accuracy and completeness of the answers. Using our synthetic data, we significantly improve the performance of the previous best list QA models by exact-match F1 scores of 5.0 on MultiSpanQA, 1.9 on Quoref, and 2.8 averaged across three BioASQ benchmarks.
comment: AAAI 2023
☆ Show me your NFT and I tell you how it will perform: Multimodal representation learning for NFT selling price prediction
Non-Fungible Tokens (NFTs) represent deeds of ownership, based on blockchain technologies and smart contracts, of unique crypto assets on digital art forms (e.g., artworks or collectibles). In the spotlight after skyrocketing in 2021, NFTs have attracted the attention of crypto enthusiasts and investors intent on placing promising investments in this profitable market. However, the NFT financial performance prediction has not been widely explored to date. In this work, we address the above problem based on the hypothesis that NFT images and their textual descriptions are essential proxies to predict the NFT selling prices. To this purpose, we propose MERLIN, a novel multimodal deep learning framework designed to train Transformer-based language and visual models, along with graph neural network models, on collections of NFTs' images and texts. A key aspect in MERLIN is its independence on financial features, as it exploits only the primary data a user interested in NFT trading would like to deal with, i.e., NFT images and textual descriptions. By learning dense representations of such data, a price-category classification task is performed by MERLIN models, which can also be tuned according to user preferences in the inference phase to mimic different risk-return investment profiles. Experimental evaluation on a publicly available dataset has shown that MERLIN models achieve significant performances according to several financial assessment criteria, fostering profitable investments, and also beating baseline machine-learning classifiers based on financial features.
comment: Accepted paper at The ACM Web Conference 2023, April 30--May 04, 2023, Austin, Texas, USA
☆ Modeling Sequential Sentence Relation to Improve Cross-lingual Dense Retrieval ICLR 2023
Recently multi-lingual pre-trained language models (PLM) such as mBERT and XLM-R have achieved impressive strides in cross-lingual dense retrieval. Despite its successes, they are general-purpose PLM while the multilingual PLM tailored for cross-lingual retrieval is still unexplored. Motivated by an observation that the sentences in parallel documents are approximately in the same order, which is universal across languages, we propose to model this sequential sentence relation to facilitate cross-lingual representation learning. Specifically, we propose a multilingual PLM called masked sentence model (MSM), which consists of a sentence encoder to generate the sentence representations, and a document encoder applied to a sequence of sentence vectors from a document. The document encoder is shared for all languages to model the universal sequential sentence relation across languages. To train the model, we propose a masked sentence prediction task, which masks and predicts the sentence vector via a hierarchical contrastive loss with sampled negatives. Comprehensive experiments on four cross-lingual retrieval tasks show MSM significantly outperforms existing advanced pre-training models, demonstrating the effectiveness and stronger cross-lingual retrieval capabilities of our approach. Code and model will be available.
comment: Published at ICLR 2023
☆ Around the world in 60 words: A generative vocabulary test for online research
Conducting experiments with diverse participants in their native languages can uncover insights into culture, cognition, and language that may not be revealed otherwise. However, conducting these experiments online makes it difficult to validate self-reported language proficiency. Furthermore, existing proficiency tests are small and cover only a few languages. We present an automated pipeline to generate vocabulary tests using text from Wikipedia. Our pipeline samples rare nouns and creates pseudowords with the same low-level statistics. Six behavioral experiments (N=236) in six countries and eight languages show that (a) our test can distinguish between native speakers of closely related languages, (b) the test is reliable ($r=0.82$), and (c) performance strongly correlates with existing tests (LexTale) and self-reports. We further show that test accuracy is negatively correlated with the linguistic distance between the tested and the native language. Our test, available in eight languages, can easily be extended to other languages.
☆ Bioformer: an efficient transformer language model for biomedical text mining
Pretrained language models such as Bidirectional Encoder Representations from Transformers (BERT) have achieved state-of-the-art performance in natural language processing (NLP) tasks. Recently, BERT has been adapted to the biomedical domain. Despite the effectiveness, these models have hundreds of millions of parameters and are computationally expensive when applied to large-scale NLP applications. We hypothesized that the number of parameters of the original BERT can be dramatically reduced with minor impact on performance. In this study, we present Bioformer, a compact BERT model for biomedical text mining. We pretrained two Bioformer models (named Bioformer8L and Bioformer16L) which reduced the model size by 60% compared to BERTBase. Bioformer uses a biomedical vocabulary and was pre-trained from scratch on PubMed abstracts and PubMed Central full-text articles. We thoroughly evaluated the performance of Bioformer as well as existing biomedical BERT models including BioBERT and PubMedBERT on 15 benchmark datasets of four different biomedical NLP tasks: named entity recognition, relation extraction, question answering and document classification. The results show that with 60% fewer parameters, Bioformer16L is only 0.1% less accurate than PubMedBERT while Bioformer8L is 0.9% less accurate than PubMedBERT. Both Bioformer16L and Bioformer8L outperformed BioBERTBase-v1.1. In addition, Bioformer16L and Bioformer8L are 2-3 fold as fast as PubMedBERT/BioBERTBase-v1.1. Bioformer has been successfully deployed to PubTator Central providing gene annotations over 35 million PubMed abstracts and 5 million PubMed Central full-text articles. We make Bioformer publicly available via https://github.com/WGLab/bioformer, including pre-trained models, datasets, and instructions for downstream use.
☆ Controlling for Stereotypes in Multimodal Language Model Evaluation
We propose a methodology and design two benchmark sets for measuring to what extent language-and-vision language models use the visual signal in the presence or absence of stereotypes. The first benchmark is designed to test for stereotypical colors of common objects, while the second benchmark considers gender stereotypes. The key idea is to compare predictions when the image conforms to the stereotype to predictions when it does not. Our results show that there is significant variation among multimodal models: the recent Transformer-based FLAVA seems to be more sensitive to the choice of image and less affected by stereotypes than older CNN-based models such as VisualBERT and LXMERT. This effect is more discernible in this type of controlled setting than in traditional evaluations where we do not know whether the model relied on the stereotype or the visual signal.
☆ Witgenstein's influence on artificial intelligence
We examine how much of the contemporary progress in artificial intelligence (and, specifically, in natural language processing), can be, more or less directly, traced back to the seminal work and ideas of the Austrian-British philosopher Ludwig Wittgenstein, with particular focus on his late views. Discussing Wittgenstein's original theses will give us the chance to survey the state of artificial intelligence, and comment on both its strengths and weaknesses. A similar text appeared first in Spanish as a chapter of CENTENARIO DEL SILENCIO (2021), a book celebrating 100 years since the publication of the Tractatus.
comment: English pre-print of a Chapter first appeared in Spanish in CENTENARIO DEL SILENCIO (2021)
☆ Using natural language processing and structured medical data to phenotype patients hospitalized due to COVID-19
To identify patients who are hospitalized because of COVID-19 as opposed to those who were admitted for other indications, we compared the performance of different computable phenotype definitions for COVID-19 hospitalizations that use different types of data from the electronic health records (EHR), including structured EHR data elements, provider notes, or a combination of both data types. And conduct a retrospective data analysis utilizing chart review-based validation. Participants are 586 hospitalized individuals who tested positive for SARS-CoV-2 during January 2022. We used natural language processing to incorporate data from provider notes and LASSO regression and Random Forests to fit classification algorithms that incorporated structured EHR data elements, provider notes, or a combination of structured data and provider notes. Results: Based on a chart review, 38% of 586 patients were determined to be hospitalized for reasons other than COVID-19 despite having tested positive for SARS-CoV-2. A classification algorithm that used provider notes had significantly better discrimination than one that used structured EHR data elements (AUROC: 0.894 vs 0.841, p < 0.001), and performed similarly to a model that combined provider notes with structured data elements (AUROC: 0.894 vs 0.893). Assessments of hospital outcome metrics significantly differed based on whether the population included all hospitalized patients who tested positive for SARS-CoV-2 versus those who were determined to have been hospitalized due to COVID-19. This work demonstrates the utility of natural language processing approaches to derive information related to patient hospitalizations in cases where there may be multiple conditions that could serve as the primary indication for hospitalization.
comment: 21 pages, 2 figures, 3 tables, 1 supplemental figure, 2 supplemental tables
☆ Revisiting Intermediate Layer Distillation for Compressing Language Models: An Overfitting Perspective
Knowledge distillation (KD) is a highly promising method for mitigating the computational problems of pre-trained language models (PLMs). Among various KD approaches, Intermediate Layer Distillation (ILD) has been a de facto standard KD method with its performance efficacy in the NLP field. In this paper, we find that existing ILD methods are prone to overfitting to training datasets, although these methods transfer more information than the original KD. Next, we present the simple observations to mitigate the overfitting of ILD: distilling only the last Transformer layer and conducting ILD on supplementary tasks. Based on our two findings, we propose a simple yet effective consistency-regularized ILD (CR-ILD), which prevents the student model from overfitting the training dataset. Substantial experiments on distilling BERT on the GLUE benchmark and several synthetic datasets demonstrate that our proposed ILD method outperforms other KD techniques. Our code is available at https://github.com/jongwooko/CR-ILD.
comment: The 17th Conference of the European Chapter of the Association for Computational Linguistics (Findings)
☆ Efficient Domain Adaptation for Speech Foundation Models
Foundation models (FMs), that are trained on broad data at scale and are adaptable to a wide range of downstream tasks, have brought large interest in the research community. Benefiting from the diverse data sources such as different modalities, languages and application domains, foundation models have demonstrated strong generalization and knowledge transfer capabilities. In this paper, we present a pioneering study towards building an efficient solution for FM-based speech recognition systems. We adopt the recently developed self-supervised BEST-RQ for pretraining, and propose the joint finetuning with both source and unsupervised target domain data using JUST Hydra. The FM encoder adapter and decoder are then finetuned to the target domain with a small amount of supervised in-domain data. On a large-scale YouTube and Voice Search task, our method is shown to be both data and model parameter efficient. It achieves the same quality with only 21.6M supervised in-domain data and 130.8M finetuned parameters, compared to the 731.1M model trained from scratch on additional 300M supervised in-domain data.
☆ Towards Few-Shot Identification of Morality Frames using In-Context Learning EMNLP 2022
Data scarcity is a common problem in NLP, especially when the annotation pertains to nuanced socio-linguistic concepts that require specialized knowledge. As a result, few-shot identification of these concepts is desirable. Few-shot in-context learning using pre-trained Large Language Models (LLMs) has been recently applied successfully in many NLP tasks. In this paper, we study few-shot identification of a psycho-linguistic concept, Morality Frames (Roy et al., 2021), using LLMs. Morality frames are a representation framework that provides a holistic view of the moral sentiment expressed in text, identifying the relevant moral foundation (Haidt and Graham, 2007) and at a finer level of granularity, the moral sentiment expressed towards the entities mentioned in the text. Previous studies relied on human annotation to identify morality frames in text which is expensive. In this paper, we propose prompting-based approaches using pretrained Large Language Models for identification of morality frames, relying only on few-shot exemplars. We compare our models' performance with few-shot RoBERTa and found promising results.
comment: Accepted to the 5th Workshop on NLP and CSS at EMNLP 2022
☆ TextShield: Beyond Successfully Detecting Adversarial Sentences in Text Classification
Adversarial attack serves as a major challenge for neural network models in NLP, which precludes the model's deployment in safety-critical applications. A recent line of work, detection-based defense, aims to distinguish adversarial sentences from benign ones. However, {the core limitation of previous detection methods is being incapable of giving correct predictions on adversarial sentences unlike defense methods from other paradigms.} To solve this issue, this paper proposes TextShield: (1) we discover a link between text attack and saliency information, and then we propose a saliency-based detector, which can effectively detect whether an input sentence is adversarial or not. (2) We design a saliency-based corrector, which converts the detected adversary sentences to benign ones. By combining the saliency-based detector and corrector, TextShield extends the detection-only paradigm to a detection-correction paradigm, thus filling the gap in the existing detection-based defense. Comprehensive experiments show that (a) TextShield consistently achieves higher or comparable performance than state-of-the-art defense methods across various attacks on different benchmarks. (b) our saliency-based detector outperforms existing detectors for detecting adversarial sentences.
☆ Improving Interpretability via Explicit Word Interaction Graph Layer AAAI 2023
Recent NLP literature has seen growing interest in improving model interpretability. Along this direction, we propose a trainable neural network layer that learns a global interaction graph between words and then selects more informative words using the learned word interactions. Our layer, we call WIGRAPH, can plug into any neural network-based NLP text classifiers right after its word embedding layer. Across multiple SOTA NLP models and various NLP datasets, we demonstrate that adding the WIGRAPH layer substantially improves NLP models' interpretability and enhances models' prediction performance at the same time.
comment: 15 pages, AAAI 2023
☆ Witscript: A System for Generating Improvised Jokes in a Conversation
A chatbot is perceived as more humanlike and likeable if it includes some jokes in its output. But most existing joke generators were not designed to be integrated into chatbots. This paper presents Witscript, a novel joke generation system that can improvise original, contextually relevant jokes, such as humorous responses during a conversation. The system is based on joke writing algorithms created by an expert comedy writer. Witscript employs well-known tools of natural language processing to extract keywords from a topic sentence and, using wordplay, to link those keywords and related words to create a punch line. Then a pretrained neural network language model that has been fine-tuned on a dataset of TV show monologue jokes is used to complete the joke response by filling the gap between the topic sentence and the punch line. A method of internal scoring filters out jokes that don't meet a preset standard of quality. Human evaluators judged Witscript's responses to input sentences to be jokes more than 40% of the time. This is evidence that Witscript represents an important next step toward giving a chatbot a humanlike sense of humor.
comment: 10 pages. Published in the Proceedings of the 12th International Conference on Computational Creativity (ICCC 2021), pages 22-31
☆ PSST! Prosodic Speech Segmentation with Transformers
Self-attention mechanisms have enabled transformers to achieve superhuman-level performance on many speech-to-text (STT) tasks, yet the challenge of automatic prosodic segmentation has remained unsolved. In this paper we finetune Whisper, a pretrained STT model, to annotate intonation unit (IU) boundaries by repurposing low-frequency tokens. Our approach achieves an accuracy of 95.8%, outperforming previous methods without the need for large-scale labeled data or enterprise grade compute resources. We also diminish input signals by applying a series of filters, finding that low pass filters at a 3.2 kHz level improve segmentation performance in out of sample and out of distribution contexts. We release our model as both a transcription tool and a baseline for further improvements in prosodic segmentation.
comment: 5 pages, 3 figures. For associated repository, see https://github.com/Nathan-Roll1/psst
☆ Measuring The Impact Of Programming Language Distribution
Current benchmarks for evaluating neural code models focus on only a small subset of programming languages, excluding many popular languages such as Go or Rust. To ameliorate this issue, we present the BabelCode framework for execution-based evaluation of any benchmark in any language. BabelCode enables new investigations into the qualitative performance of models' memory, runtime, and individual test case results. Additionally, we present a new code translation dataset called Translating Python Programming Puzzles (TP3) from the Python Programming Puzzles (Schuster et al. 2021) benchmark that involves translating expert-level python functions to any language. With both BabelCode and the TP3 benchmark, we investigate if balancing the distributions of 14 languages in a training dataset improves a large language model's performance on low-resource languages. Training a model on a balanced corpus results in, on average, 12.34% higher $pass@k$ across all tasks and languages compared to the baseline. We find that this strategy achieves 66.48% better $pass@k$ on low-resource languages at the cost of only a 12.94% decrease to high-resource languages. In our three translation tasks, this strategy yields, on average, 30.77% better low-resource $pass@k$ while having 19.58% worse high-resource $pass@k$.
comment: Code and data release: https://github.com/google-research/babelcode
☆ LaMPP: Language Models as Probabilistic Priors for Perception and Action
Language models trained on large text corpora encode rich distributional information about real-world environments and action sequences. This information plays a crucial role in current approaches to language processing tasks like question answering and instruction generation. We describe how to leverage language models for *non-linguistic* perception and control tasks. Our approach casts labeling and decision-making as inference in probabilistic graphical models in which language models parameterize prior distributions over labels, decisions and parameters, making it possible to integrate uncertain observations and incomplete background knowledge in a principled way. Applied to semantic segmentation, household navigation, and activity recognition tasks, this approach improves predictions on rare, out-of-distribution, and structurally novel inputs.
comment: 12 pages, 4 tables, 4 figures
☆ CAB: Empathetic Dialogue Generation with Cognition, Affection and Behavior DASFAA 2023
Empathy is an important characteristic to be considered when building a more intelligent and humanized dialogue agent. However, existing methods did not fully comprehend empathy as a complex process involving three aspects: cognition, affection and behavior. In this paper, we propose CAB, a novel framework that takes a comprehensive perspective of cognition, affection and behavior to generate empathetic responses. For cognition, we build paths between critical keywords in the dialogue by leveraging external knowledge. This is because keywords in a dialogue are the core of sentences. Building the logic relationship between keywords, which is overlooked by the majority of existing works, can improve the understanding of keywords and contextual logic, thus enhance the cognitive ability. For affection, we capture the emotional dependencies with dual latent variables that contain both interlocutors' emotions. The reason is that considering both interlocutors' emotions simultaneously helps to learn the emotional dependencies. For behavior, we use appropriate dialogue acts to guide the dialogue generation to enhance the empathy expression. Extensive experiments demonstrate that our multi-perspective model outperforms the state-of-the-art models in both automatic and manual evaluation.
comment: accepted as a short paper at DASFAA 2023
☆ Detecting Reddit Users with Depression Using a Hybrid Neural Network
Depression is a widespread mental health issue, affecting an estimated 3.8% of the global population. It is also one of the main contributors to disability worldwide. Recently it is becoming popular for individuals to use social media platforms (e.g., Reddit) to express their difficulties and health issues (e.g., depression) and seek support from other users in online communities. It opens great opportunities to automatically identify social media users with depression by parsing millions of posts for potential interventions. Deep learning methods have begun to dominate in the field of machine learning and natural language processing (NLP) because of their ease of use, efficient processing, and state-of-the-art results on many NLP tasks. In this work, we propose a hybrid deep learning model which combines a pretrained sentence BERT (SBERT) and convolutional neural network (CNN) to detect individuals with depression with their Reddit posts. The sentence BERT is used to learn the meaningful representation of semantic information in each post. CNN enables the further transformation of those embeddings and the temporal identification of behavioral patterns of users. We trained and evaluated the model performance to identify Reddit users with depression by utilizing the Self-reported Mental Health Diagnoses (SMHD) data. The hybrid deep learning model achieved an accuracy of 0.86 and an F1 score of 0.86 and outperformed the state-of-the-art documented result (F1 score of 0.79) by other machine learning models in the literature. The results show the feasibility of the hybrid model to identify individuals with depression. Although the hybrid model is validated to detect depression with Reddit posts, it can be easily tuned and applied to other text classification tasks and different clinical applications.
☆ Witscript 2: A System for Generating Improvised Jokes Without Wordplay
A previous paper presented Witscript, a system for generating conversational jokes that rely on wordplay. This paper extends that work by presenting Witscript 2, which uses a large language model to generate conversational jokes that rely on common sense instead of wordplay. Like Witscript, Witscript 2 is based on joke-writing algorithms created by an expert comedy writer. Human evaluators judged Witscript 2's responses to input sentences to be jokes 46% of the time, compared to 70% of the time for human-written responses. This is evidence that Witscript 2 represents another step toward giving a chatbot a humanlike sense of humor.
comment: 5 pages. Published in the Proceedings of the 13th International Conference on Computational Creativity (ICCC 2022), pages 54-58. arXiv admin note: text overlap with arXiv:2301.02695. substantial text overlap with arXiv:2302.02008
♻ ☆ Vicarious Offense and Noise Audit of Offensive Speech Classifiers
This paper examines social web content moderation from two key perspectives: automated methods (machine moderators) and human evaluators (human moderators). We conduct a noise audit at an unprecedented scale using nine machine moderators trained on well-known offensive speech data sets evaluated on a corpus sampled from 92 million YouTube comments discussing a multitude of issues relevant to US politics. We introduce a first-of-its-kind data set of vicarious offense. We ask annotators: (1) if they find a given social media post offensive; and (2) how offensive annotators sharing different political beliefs would find the same content. Our experiments with machine moderators reveal that moderation outcomes wildly vary across different machine moderators. Our experiments with human moderators suggest that (1) political leanings considerably affect first-person offense perspective; (2) Republicans are the worst predictors of vicarious offense; (3) predicting vicarious offense for the Republicans is most challenging than predicting vicarious offense for the Independents and the Democrats; and (4) disagreement across political identity groups considerably increases when sensitive issues such as reproductive rights or gun control/rights are discussed. Both experiments suggest that offense, is indeed, highly subjective and raise important questions concerning content moderation practices.
♻ ☆ TopoBERT: Plug and Play Toponym Recognition Module Harnessing Fine-tuned BERT
Extracting precise geographical information from textual contents is crucial in a plethora of applications. For example, during hazardous events, a robust and unbiased toponym extraction framework can provide an avenue to tie the location concerned to the topic discussed by news media posts and pinpoint humanitarian help requests or damage reports from social media. Early studies have leveraged rule-based, gazetteer-based, deep learning, and hybrid approaches to address this problem. However, the performance of existing tools is deficient in supporting operations like emergency rescue, which relies on fine-grained, accurate geographic information. The emerging pretrained language models can better capture the underlying characteristics of text information, including place names, offering a promising pathway to optimize toponym recognition to underpin practical applications. In this paper, TopoBERT, a toponym recognition module based on a one dimensional Convolutional Neural Network (CNN1D) and Bidirectional Encoder Representation from Transformers (BERT), is proposed and fine-tuned. Three datasets (CoNLL2003-Train, Wikipedia3000, WNUT2017) are leveraged to tune the hyperparameters, discover the best training strategy, and train the model. Another two datasets (CoNLL2003-Test and Harvey2017) are used to evaluate the performance. Three distinguished classifiers, linear, multi-layer perceptron, and CNN1D, are benchmarked to determine the optimal model architecture. TopoBERT achieves state-of-the-art performance (f1-score=0.865) compared to the other five baseline models and can be applied to diverse toponym recognition tasks without additional training.
comment: 8 Pages, 6 figures
♻ ☆ SemEval 2023 Task 9: Multilingual Tweet Intimacy Analysis SemEval 2023
We propose MINT, a new Multilingual INTimacy analysis dataset covering 13,372 tweets in 10 languages including English, French, Spanish, Italian, Portuguese, Korean, Dutch, Chinese, Hindi, and Arabic. We benchmarked a list of popular multilingual pre-trained language models. The dataset is released along with the SemEval 2023 Task 9: Multilingual Tweet Intimacy Analysis (https://sites.google.com/umich.edu/semeval-2023-tweet-intimacy).
comment: SemEval 2023 Task 9: Multilingual Tweet Intimacy Analysis
♻ ☆ Resolving Open-textured Rules with Templated Interpretive Arguments
Open-textured terms in written rules are typically settled through interpretive argumentation. Ongoing work has attempted to catalogue the schemes used in such interpretive argumentation. But how can the use of these schemes affect the way in which people actually use and reason over the proper interpretations of open-textured terms? Using the interpretive argument-eliciting game Aporia as our framework, we carried out an empirical study to answer this question. Differing from previous work, we did not allow participants to argue for interpretations arbitrarily, but to only use arguments that fit with a given set of interpretive argument templates. Finally, we analyze the results captured by this new dataset, specifically focusing on practical implications for the development of interpretation-capable artificial reasoners.
comment: Presented at the 2022 European Conference on Argumentation (ECA)
♻ ☆ AmbiCoref: Evaluating Human and Model Sensitivity to Ambiguous Coreference EACL 2023
Given a sentence "Abby told Brittney that she upset Courtney", one would struggle to understand who "she" refers to, and ask for clarification. However, if the word "upset" were replaced with "hugged", "she" unambiguously refers to Abby. We study if modern coreference resolution models are sensitive to such pronominal ambiguity. To this end, we construct AmbiCoref, a diagnostic corpus of minimal sentence pairs with ambiguous and unambiguous referents. Our examples generalize psycholinguistic studies of human perception of ambiguity around particular arrangements of verbs and their arguments. Analysis shows that (1) humans are less sure of referents in ambiguous AmbiCoref examples than unambiguous ones, and (2) most coreference models show little difference in output between ambiguous and unambiguous pairs. We release AmbiCoref as a diagnostic corpus for testing whether models treat ambiguity similarly to humans.
comment: EACL 2023 Findings
♻ ☆ CoNT: Contrastive Neural Text Generation NeurIPS 2022
Recently, contrastive learning attracts increasing interests in neural text generation as a new solution to alleviate the exposure bias problem. It introduces a sequence-level training signal which is crucial to generation tasks that always rely on auto-regressive decoding. However, previous methods using contrastive learning in neural text generation usually lead to inferior performance. In this paper, we analyse the underlying reasons and propose a new Contrastive Neural Text generation framework, CoNT. CoNT addresses bottlenecks that prevent contrastive learning from being widely adopted in generation tasks from three aspects -- the construction of contrastive examples, the choice of the contrastive loss, and the strategy in decoding. We validate CoNT on five generation tasks with ten benchmarks, including machine translation, summarization, code comment generation, data-to-text generation and commonsense generation. Experimental results show that CoNT clearly outperforms the conventional training framework on all the ten benchmarks with a convincing margin. Especially, CoNT surpasses previous the most competitive contrastive learning method for text generation, by 1.50 BLEU on machine translation and 1.77 ROUGE-1 on summarization, respectively. It achieves new state-of-the-art on summarization, code comment generation (without external data) and data-to-text generation.
comment: Accepted by NeurIPS 2022
♻ ☆ MetaPrompting: Learning to Learn Better Prompts COLING 2022
Prompting method is regarded as one of the crucial progress for few-shot nature language processing. Recent research on prompting moves from discrete tokens based ``hard prompts'' to continuous ``soft prompts'', which employ learnable vectors as pseudo prompt tokens and achieve better performance. Though showing promising prospects, these soft-prompting methods are observed to rely heavily on good initialization to take effect. Unfortunately, obtaining a perfect initialization for soft prompts requires understanding of inner language models working and elaborate design, which is no easy task and has to restart from scratch for each new task. To remedy this, we propose a generalized soft prompting method called MetaPrompting, which adopts the well-recognized model-agnostic meta-learning algorithm to automatically find better prompt initialization that facilitates fast adaptation to new prompting tasks.Extensive experiments show MetaPrompting tackles soft prompt initialization problem and brings significant improvement on four different datasets (over 6 points improvement in accuracy for 1-shot setting), achieving new state-of-the-art performance.
comment: Accepted as COLING 2022 long paper
♻ ☆ Learning the Effects of Physical Actions in a Multi-modal Environment
Large Language Models (LLMs) handle physical commonsense information inadequately. As a result of being trained in a disembodied setting, LLMs often fail to predict an action's outcome in a given environment. However, predicting the effects of an action before it is executed is crucial in planning, where coherent sequences of actions are often needed to achieve a goal. Therefore, we introduce the multi-modal task of predicting the outcomes of actions solely from realistic sensory inputs (images and text). Next, we extend an LLM to model latent representations of objects to better predict action outcomes in an environment. We show that multi-modal models can capture physical commonsense when augmented with visual information. Finally, we evaluate our model's performance on novel actions and objects and find that combining modalities help models to generalize and learn physical commonsense reasoning better.
♻ ☆ Leveraging task dependency and contrastive learning for Legal Judgement Prediction on the European Court of Human Rights EACL 2023
We report on an experiment in legal judgement prediction on European Court of Human Rights cases where our model first learns to predict the convention articles allegedly violated by the state from case facts descriptions, and subsequently utilizes that information to predict a finding of a violation by the court. We assess the dependency between these two tasks at the feature and outcome level. Furthermore, we leverage a hierarchical contrastive loss to pull together article specific representations of cases at the higher level level, leading to distinctive article clusters, and further pulls the cases in each article cluster based on their outcome leading to sub-clusters of cases with similar outcomes. Our experiment results demonstrate that, given a static pre-trained encoder, our models produce a small but consistent improvement in prediction performance over single-task and joint models without contrastive loss.
comment: Accepted to EACL 2023
♻ ☆ Zero Shot Transfer of Legal Judgement Prediction as Article-aware Entailment for the European Court of Human Rights EACL
In this paper, we cast Legal Judgment Prediction (LJP) from text on European Court of Human Rights cases as an entailment task, where the case outcome is classified from a combined input of case facts and convention articles. This configuration facilitates the model learning legal reasoning ability in mapping article text to specific fact text. It also provides the opportunity to evaluate the model's ability to generalize to zero-shot settings when asked to classify the case outcome with respect to articles not seen during training. We devise zero-shot LJP experiments and apply domain adaptation methods based on domain discriminator and Wasserstein distance. Our results demonstrate that the entailment architecture outperforms straightforward fact classification. We also find that domain adaptation methods improve zero-shot transfer performance, with article relatedness and encoder pre-training influencing the effect.
comment: Accepted to EACL Findings 2023
♻ ☆ Visualize Before You Write: Imagination-Guided Open-Ended Text Generation EACL 2023
Recent advances in text-to-image synthesis make it possible to visualize machine imaginations for a given context. On the other hand, when generating text, human writers are gifted at creative visualization, which enhances their writings by forming imaginations as blueprints before putting down the stories in words. Inspired by such a cognitive process, we ask the natural question of whether we can endow machines with the same ability to utilize visual information and construct a general picture of the context to guide text generation. In this work, we propose iNLG that uses machine-generated images to guide language models in open-ended text generation. The experiments and analyses demonstrate the effectiveness of iNLG on open-ended text generation tasks, including text completion, story generation, and concept-to-text generation in both few-shot and full-data scenarios. Both automatic metrics and human evaluations verify that the text snippets generated by our iNLG are coherent and informative while displaying minor degeneration.
comment: EACL 2023
♻ ☆ Language Quantized AutoEncoders: Towards Unsupervised Text-Image Alignment
Recent progress in scaling up large language models has shown impressive capabilities in performing few-shot learning across a wide range of text-based tasks. However, a key limitation is that these language models fundamentally lack visual perception - a crucial attribute needed to extend these models to be able to interact with the real world and solve vision tasks, such as in visual-question answering and robotics. Prior works have largely connected image to text through pretraining and/or fine-tuning on curated image-text datasets, which can be a costly and expensive process. In order to resolve this limitation, we propose a simple yet effective approach called Language-Quantized AutoEncoder (LQAE), a modification of VQ-VAE that learns to align text-image data in an unsupervised manner by leveraging pretrained language models (e.g., BERT, RoBERTa). Our main idea is to encode image as sequences of text tokens by directly quantizing image embeddings using a pretrained language codebook. We then apply random masking followed by a BERT model, and have the decoder reconstruct the original image from BERT predicted text token embeddings. By doing so, LQAE learns to represent similar images with similar clusters of text tokens, thereby aligning these two modalities without the use of aligned text-image pairs. This enables few-shot image classification with large language models (e.g., GPT-3) as well as linear classification of images based on BERT text features. To the best of our knowledge, our work is the first work that uses unaligned images for multimodal tasks by leveraging the power of pretrained language models.
comment: Fixed typos
♻ ☆ VaxxHesitancy: A Dataset for Studying Hesitancy Towards COVID-19 Vaccination on Twitter
Vaccine hesitancy has been a common concern, probably since vaccines were created and, with the popularisation of social media, people started to express their concerns about vaccines online alongside those posting pro- and anti-vaccine content. Predictably, since the first mentions of a COVID-19 vaccine, social media users posted about their fears and concerns or about their support and belief into the effectiveness of these rapidly developing vaccines. Identifying and understanding the reasons behind public hesitancy towards COVID-19 vaccines is important for policy markers that need to develop actions to better inform the population with the aim of increasing vaccine take-up. In the case of COVID-19, where the fast development of the vaccines was mirrored closely by growth in anti-vaxx disinformation, automatic means of detecting citizen attitudes towards vaccination became necessary. This is an important computational social sciences task that requires data analysis in order to gain in-depth understanding of the phenomena at hand. Annotated data is also necessary for training data-driven models for more nuanced analysis of attitudes towards vaccination. To this end, we created a new collection of over 3,101 tweets annotated with users' attitudes towards COVID-19 vaccination (stance). Besides, we also develop a domain-specific language model (VaxxBERT) that achieves the best predictive performance (73.0 accuracy and 69.3 F1-score) as compared to a robust set of baselines. To the best of our knowledge, these are the first dataset and model that model vaccine hesitancy as a category distinct from pro- and anti-vaccine stance.
♻ ☆ "John is 50 years old, can his son be 65?" Evaluating NLP Models' Understanding of Feasibility EACL 2023
In current NLP research, large-scale language models and their abilities are widely being discussed. Some recent works have also found notable failures of these models. Often these failure examples involve complex reasoning abilities. This work focuses on a simple commonsense ability, reasoning about when an action (or its effect) is feasible. To this end, we introduce FeasibilityQA, a question-answering dataset involving binary classification (BCQ) and multi-choice multi-correct questions (MCQ) that test understanding of feasibility. We show that even state-of-the-art models such as GPT-3, GPT-2, and T5 struggle to answer the feasibility questions correctly. Specifically, on MCQ and BCQ questions, GPT-3 achieves an accuracy of just (19%, 62%) and (25%, 64%) in zero-shot and few-shot settings, respectively. We also evaluate models by providing relevant knowledge statements required to answer the question. We find that the additional knowledge leads to a 7% gain in performance, but the overall performance still remains low. These results make one wonder how much commonsense knowledge about action feasibility is encoded in state-of-the-art models and how well they can reason about it.
comment: EACL 2023
♻ ☆ The Solvability of Interpretability Evaluation Metrics EACL 2023
Feature attribution methods are popular for explaining neural network predictions, and they are often evaluated on metrics such as comprehensiveness and sufficiency. In this paper, we highlight an intriguing property of these metrics: their solvability. Concretely, we can define the problem of optimizing an explanation for a metric, which can be solved by beam search. This observation leads to the obvious yet unaddressed question: why do we use explainers (e.g., LIME) not based on solving the target metric, if the metric value represents explanation quality? We present a series of investigations showing strong performance of this beam search explainer and discuss its broader implication: a definition-evaluation duality of interpretability concepts. We implement the explainer and release the Python solvex package for models of text, image and tabular domains.
comment: EACL 2023 (Findings). Project website at https://yilunzhou.github.io/solvability/
♻ ☆ A Survey of Active Learning for Natural Language Processing EMNLP 2022
In this work, we provide a survey of active learning (AL) for its applications in natural language processing (NLP). In addition to a fine-grained categorization of query strategies, we also investigate several other important aspects of applying AL to NLP problems. These include AL for structured prediction tasks, annotation cost, model learning (especially with deep neural models), and starting and stopping AL. Finally, we conclude with a discussion of related topics and future directions.
comment: EMNLP 2022
♻ ☆ Keyword Assisted Topic Models
In recent years, fully automated content analysis based on probabilistic topic models has become popular among social scientists because of their scalability. The unsupervised nature of the models makes them suitable for exploring topics in a corpus without prior knowledge. However, researchers find that these models often fail to measure specific concepts of substantive interest by inadvertently creating multiple topics with similar content and combining distinct themes into a single topic. In this paper, we empirically demonstrate that providing a small number of keywords can substantially enhance the measurement performance of topic models. An important advantage of the proposed keyword assisted topic model (keyATM) is that the specification of keywords requires researchers to label topics prior to fitting a model to the data. This contrasts with a widespread practice of post-hoc topic interpretation and adjustments that compromises the objectivity of empirical findings. In our application, we find that keyATM provides more interpretable results, has better document classification performance, and is less sensitive to the number of topics than the standard topic models. Finally, we show that keyATM can also incorporate covariates and model time trends. An open-source software package is available for implementing the proposed methodology.
♻ ☆ Logically at Factify 2: A Multi-Modal Fact Checking System Based on Evidence Retrieval techniques and Transformer Encoder Architecture AAAI'23
In this paper, we present the Logically submissions to De-Factify 2 challenge (DE-FACTIFY 2023) on the task 1 of Multi-Modal Fact Checking. We describes our submissions to this challenge including explored evidence retrieval and selection techniques, pre-trained cross-modal and unimodal models, and a cross-modal veracity model based on the well established Transformer Encoder (TE) architecture which is heavily relies on the concept of self-attention. Exploratory analysis is also conducted on this Factify 2 data set that uncovers the salient multi-modal patterns and hypothesis motivating the architecture proposed in this work. A series of preliminary experiments were done to investigate and benchmarking different pre-trained embedding models, evidence retrieval settings and thresholds. The final system, a standard two-stage evidence based veracity detection system, yields weighted avg. 0.79 on both val set and final blind test set on the task 1, which achieves 3rd place with a small margin to the top performing system on the leaderboard among 9 participants.
comment: Accepted in AAAI'23: Second Workshop on Multimodal Fact-Checking and Hate Speech Detection, February 2023, Washington, DC, USA
♻ ☆ A Discerning Several Thousand Judgments: GPT-3 Rates the Article + Adjective + Numeral + Noun Construction
Knowledge of syntax includes knowledge of rare, idiosyncratic constructions. LLMs must overcome frequency biases in order to master such constructions. In this study, I prompt GPT-3 to give acceptability judgments on the English-language Article + Adjective + Numeral + Noun construction (e.g., "a lovely five days"). I validate the prompt using the CoLA corpus of acceptability judgments and then zero in on the AANN construction. I compare GPT- 3's judgments to crowdsourced human judgments on a subset of sentences. GPT-3's judgments are broadly similar to human judgments and generally align with proposed constraints in the literature but, in some cases, GPT-3's judgments and human judgments diverge from the literature and from each other.
♻ ☆ Learning Automata-Based Task Knowledge Representation from Large-Scale Generative Language Models
Automata-based representations play an important role in control and planning in sequential decision-making, but obtaining high-level task knowledge for building automata is often difficult. Although large-scale generative language models (GLMs) can help automatically distill task knowledge, the textual outputs from GLMs are not amenable for formal verification or use in sequential decision-making. We propose a novel algorithm named GLM2FSA, which obtains high-level task knowledge represented in a finite state automaton (FSA) from a given brief description of the task goal. GLM2FSA sends queries to a GLM for task knowledge in textual form and then builds an FSA to represent the textual knowledge. It fills the gap between text and automata-based representations, and the constructed FSA can be directly utilized in formal verification. We provide an algorithm for iteratively refining the queries to the GLM based on the outcomes, e.g., counter-examples, from verification. We demonstrate the algorithm on examples that range from everyday tasks, e.g., crossing a road and making coffee, to security applications to laboratory safety protocols.
Computer Vision and Pattern Recognition 59
☆ Egocentric Video Task Translation @ Ego4D Challenge 2022 ECCV
This technical report describes the EgoTask Translation approach that explores relations among a set of egocentric video tasks in the Ego4D challenge. To improve the primary task of interest, we propose to leverage existing models developed for other related tasks and design a task translator that learns to ''translate'' auxiliary task features to the primary task. With no modification to the baseline architectures, our proposed approach achieves competitive performance on two Ego4D challenges, ranking the 1st in the talking to me challenge and the 3rd in the PNR keyframe localization challenge.
comment: The technical report of ECCV@2022 Ego4D challenge
☆ Enhancing Once-For-All: A Study on Parallel Blocks, Skip Connections and Early Exits
The use of Neural Architecture Search (NAS) techniques to automate the design of neural networks has become increasingly popular in recent years. The proliferation of devices with different hardware characteristics using such neural networks, as well as the need to reduce the power consumption for their search, has led to the realisation of Once-For-All (OFA), an eco-friendly algorithm characterised by the ability to generate easily adaptable models through a single learning process. In order to improve this paradigm and develop high-performance yet eco-friendly NAS techniques, this paper presents OFAv2, the extension of OFA aimed at improving its performance while maintaining the same ecological advantage. The algorithm is improved from an architectural point of view by including early exits, parallel blocks and dense skip connections. The training process is extended by two new phases called Elastic Level and Elastic Height. A new Knowledge Distillation technique is presented to handle multi-output networks, and finally a new strategy for dynamic teacher network selection is proposed. These modifications allow OFAv2 to improve its accuracy performance on the Tiny ImageNet dataset by up to 12.07% compared to the original version of OFA, while maintaining the algorithm flexibility and advantages.
☆ IKEA-Manual: Seeing Shape Assembly Step by Step NeurIPS 2022
Human-designed visual manuals are crucial components in shape assembly activities. They provide step-by-step guidance on how we should move and connect different parts in a convenient and physically-realizable way. While there has been an ongoing effort in building agents that perform assembly tasks, the information in human-design manuals has been largely overlooked. We identify that this is due to 1) a lack of realistic 3D assembly objects that have paired manuals and 2) the difficulty of extracting structured information from purely image-based manuals. Motivated by this observation, we present IKEA-Manual, a dataset consisting of 102 IKEA objects paired with assembly manuals. We provide fine-grained annotations on the IKEA objects and assembly manuals, including decomposed assembly parts, assembly plans, manual segmentation, and 2D-3D correspondence between 3D parts and visual manuals. We illustrate the broad application of our dataset on four tasks related to shape assembly: assembly plan generation, part segmentation, pose estimation, and 3D part assembly.
comment: NeurIPS 2022 Datasets and Benchmarks Track. Project page: https://cs.stanford.edu/~rcwang/projects/ikea_manual
☆ MOSE: A New Dataset for Video Object Segmentation in Complex Scenes
Video object segmentation (VOS) aims at segmenting a particular object throughout the entire video clip sequence. The state-of-the-art VOS methods have achieved excellent performance (e.g., 90+% J&F) on existing datasets. However, since the target objects in these existing datasets are usually relatively salient, dominant, and isolated, VOS under complex scenes has rarely been studied. To revisit VOS and make it more applicable in the real world, we collect a new VOS dataset called coMplex video Object SEgmentation (MOSE) to study the tracking and segmenting objects in complex environments. MOSE contains 2,149 video clips and 5,200 objects from 36 categories, with 431,725 high-quality object segmentation masks. The most notable feature of MOSE dataset is complex scenes with crowded and occluded objects. The target objects in the videos are commonly occluded by others and disappear in some frames. To analyze the proposed MOSE dataset, we benchmark 18 existing VOS methods under 4 different settings on the proposed MOSE dataset and conduct comprehensive comparisons. The experiments show that current VOS algorithms cannot well perceive objects in complex scenes. For example, under the semi-supervised VOS setting, the highest J&F by existing state-of-the-art VOS methods is only 59.4% on MOSE, much lower than their ~90% J&F performance on DAVIS. The results reveal that although excellent performance has been achieved on existing benchmarks, there are unresolved challenges under complex scenes and more efforts are desired to explore these challenges in the future. The proposed MOSE dataset has been released at https://henghuiding.github.io/MOSE.
comment: MOSE Dataset Report
☆ MorDIFF: Recognition Vulnerability and Attack Detectability of Face Morphing Attacks Created by Diffusion Autoencoders
Investigating new methods of creating face morphing attacks is essential to foresee novel attacks and help mitigate them. Creating morphing attacks is commonly either performed on the image-level or on the representation-level. The representation-level morphing has been performed so far based on generative adversarial networks (GAN) where the encoded images are interpolated in the latent space to produce a morphed image based on the interpolated vector. Such a process was constrained by the limited reconstruction fidelity of GAN architectures. Recent advances in the diffusion autoencoder models have overcome the GAN limitations, leading to high reconstruction fidelity. This theoretically makes them a perfect candidate to perform representation-level face morphing. This work investigates using diffusion autoencoders to create face morphing attacks by comparing them to a wide range of image-level and representation-level morphs. Our vulnerability analyses on four state-of-the-art face recognition models have shown that such models are highly vulnerable to the created attacks, the MorDIFF, especially when compared to existing representation-level morphs. Detailed detectability analyses are also performed on the MorDIFF, showing that they are as challenging to detect as other morphing attacks created on the image- or representation-level. Data and morphing script are made public.
☆ vMAP: Vectorised Object Mapping for Neural Field SLAM
We present vMAP, an object-level dense SLAM system using neural field representations. Each object is represented by a small MLP, enabling efficient, watertight object modelling without the need for 3D priors. As an RGB-D camera browses a scene with no prior information, vMAP detects object instances on-the-fly, and dynamically adds them to its map. Specifically, thanks to the power of vectorised training, vMAP can optimise as many as 50 individual objects in a single scene, with an extremely efficient training speed of 5Hz map update. We experimentally demonstrate significantly improved scene-level and object-level reconstruction quality compared to prior neural field SLAM systems.
comment: Project Page:https://kxhit.github.io/vMAP
☆ HDFormer: High-order Directed Transformer for 3D Human Pose Estimation
Human pose estimation is a complicated structured data sequence modeling task. Most existing methods only consider the pair-wise interaction of human body joints in model learning. Unfortunately, this causes 3D pose estimation to fail in difficult cases such as $\textit{joints overlapping}$, and pose $\textit{fast-changing}$, as pair-wise relations cannot exploit fine-grained human body priors in pose estimation. To this end, we revamped the 3D pose estimation framework with a $\textit{High-order}$ $\textit{Directed}$ $\textit{Transformer}$ (HDFormer), which coherently exploits the high-order bones and joints relevances to boost the performance of pose estimation. Specifically, HDFormer adopts both self-attention and high-order attention schemes to build up a multi-order attention module to perform the information flow interaction including the first-order $"\textit{joint$\leftrightarrow$joint}"$, second-order $"\textit{bone$\leftrightarrow$joint}"$ as well as high-order $"\textit{hyperbone$\leftrightarrow$joint}"$ relationships (hyperbone is defined as a joint set), compensating the hard cases prediction in fast-changing and heavy occlusion scenarios. Moreover, modernized CNN techniques are applied to upgrade the transformer-based architecture to speed up the HDFormer, achieving a favorable trade-off between effectiveness and efficiency. We compare our model with other SOTA models on the datasets Human3.6M and MPI-INF-3DHP. The results demonstrate that the proposed HDFormer achieves superior performance with only $\textbf{1/10}$ parameters and much lower computational cost compared to the current SOTAs. Moreover, HDFormer can be applied to various types of real-world applications, enabling real-time and accurate 3D pose estimation. The source code is in https://github.com/hyer/HDFormer.
☆ Leveraging weak complementary labels to improve semantic segmentation of hepatocellular carcinoma and cholangiocarcinoma in H&E-stained slides
In this paper, we present a deep learning segmentation approach to classify and quantify the two most prevalent primary liver cancers - hepatocellular carcinoma and intrahepatic cholangiocarcinoma - from hematoxylin and eosin (H&E) stained whole slide images. While semantic segmentation of medical images typically requires costly pixel-level annotations by domain experts, there often exists additional information which is routinely obtained in clinical diagnostics but rarely utilized for model training. We propose to leverage such weak information from patient diagnoses by deriving complementary labels that indicate to which class a sample cannot belong to. To integrate these labels, we formulate a complementary loss for segmentation. Motivated by the medical application, we demonstrate for general segmentation tasks that including additional patches with solely weak complementary labels during model training can significantly improve the predictive performance and robustness of a model. On the task of diagnostic differentiation between hepatocellular carcinoma and intrahepatic cholangiocarcinoma, we achieve a balanced accuracy of 0.91 (CI 95%: 0.86 - 0.95) at case level for 165 hold-out patients. Furthermore, we also show that leveraging complementary labels improves the robustness of segmentation and increases performance at case level.
Self-Supervised In-Domain Representation Learning for Remote Sensing Image Scene Classification
Transferring the ImageNet pre-trained weights to the various remote sensing tasks has produced acceptable results and reduced the need for labeled samples. However, the domain differences between ground imageries and remote sensing images cause the performance of such transfer learning to be limited. Recent research has demonstrated that self-supervised learning methods capture visual features that are more discriminative and transferable than the supervised ImageNet weights. We are motivated by these facts to pre-train the in-domain representations of remote sensing imagery using contrastive self-supervised learning and transfer the learned features to other related remote sensing datasets. Specifically, we used the SimSiam algorithm to pre-train the in-domain knowledge of remote sensing datasets and then transferred the obtained weights to the other scene classification datasets. Thus, we have obtained state-of-the-art results on five land cover classification datasets with varying numbers of classes and spatial resolutions. In addition, By conducting appropriate experiments, including feature pre-training using datasets with different attributes, we have identified the most influential factors that make a dataset a good choice for obtaining in-domain features. We have transferred the features obtained by pre-training SimSiam on remote sensing datasets to various downstream tasks and used them as initial weights for fine-tuning. Moreover, we have linearly evaluated the obtained representations in cases where the number of samples per class is limited. Our experiments have demonstrated that using a higher-resolution dataset during the self-supervised pre-training stage results in learning more discriminative and general representations.
☆ DilateFormer: Multi-Scale Dilated Transformer for Visual Recognition
As a de facto solution, the vanilla Vision Transformers (ViTs) are encouraged to model long-range dependencies between arbitrary image patches while the global attended receptive field leads to quadratic computational cost. Another branch of Vision Transformers exploits local attention inspired by CNNs, which only models the interactions between patches in small neighborhoods. Although such a solution reduces the computational cost, it naturally suffers from small attended receptive fields, which may limit the performance. In this work, we explore effective Vision Transformers to pursue a preferable trade-off between the computational complexity and size of the attended receptive field. By analyzing the patch interaction of global attention in ViTs, we observe two key properties in the shallow layers, namely locality and sparsity, indicating the redundancy of global dependency modeling in shallow layers of ViTs. Accordingly, we propose Multi-Scale Dilated Attention (MSDA) to model local and sparse patch interaction within the sliding window. With a pyramid architecture, we construct a Multi-Scale Dilated Transformer (DilateFormer) by stacking MSDA blocks at low-level stages and global multi-head self-attention blocks at high-level stages. Our experiment results show that our DilateFormer achieves state-of-the-art performance on various vision tasks. On ImageNet-1K classification task, DilateFormer achieves comparable performance with 70% fewer FLOPs compared with existing state-of-the-art models. Our DilateFormer-Base achieves 85.6% top-1 accuracy on ImageNet-1K classification task, 53.5% box mAP/46.1% mask mAP on COCO object detection/instance segmentation task and 51.1% MS mIoU on ADE20K semantic segmentation task.
comment: Accepted to IEEE Transaction on Multimedia, 2023 (Submission date: 22-Sep-2022)
☆ Understanding metric-related pitfalls in image analysis validation
Validation metrics are key for the reliable tracking of scientific progress and for bridging the current chasm between artificial intelligence (AI) research and its translation into practice. However, increasing evidence shows that particularly in image analysis, metrics are often chosen inadequately in relation to the underlying research problem. This could be attributed to a lack of accessibility of metric-related knowledge: While taking into account the individual strengths, weaknesses, and limitations of validation metrics is a critical prerequisite to making educated choices, the relevant knowledge is currently scattered and poorly accessible to individual researchers. Based on a multi-stage Delphi process conducted by a multidisciplinary expert consortium as well as extensive community feedback, the present work provides the first reliable and comprehensive common point of access to information on pitfalls related to validation metrics in image analysis. Focusing on biomedical image analysis but with the potential of transfer to other fields, the addressed pitfalls generalize across application domains and are categorized according to a newly created, domain-agnostic taxonomy. To facilitate comprehension, illustrations and specific examples accompany each pitfall. As a structured body of information accessible to researchers of all levels of expertise, this work enhances global comprehension of a key topic in image analysis validation.
☆ IMPORTANT-Net: Integrated MRI Multi-Parameter Reinforcement Fusion Generator with Attention Network for Synthesizing Absent Data
Magnetic resonance imaging (MRI) is highly sensitive for lesion detection in the breasts. Sequences obtained with different settings can capture the specific characteristics of lesions. Such multi-parameter MRI information has been shown to improve radiologist performance in lesion classification, as well as improving the performance of artificial intelligence models in various tasks. However, obtaining multi-parameter MRI makes the examination costly in both financial and time perspectives, and there may be safety concerns for special populations, thus making acquisition of the full spectrum of MRI sequences less durable. In this study, different than naive input fusion or feature concatenation from existing MRI parameters, a novel $\textbf{I}$ntegrated MRI $\textbf{M}$ulti-$\textbf{P}$arameter reinf$\textbf{O}$rcement fusion generato$\textbf{R}$ wi$\textbf{T}$h $\textbf{A}$tte$\textbf{NT}$ion Network (IMPORTANT-Net) is developed to generate missing parameters. First, the parameter reconstruction module is used to encode and restore the existing MRI parameters to obtain the corresponding latent representation information at any scale level. Then the multi-parameter fusion with attention module enables the interaction of the encoded information from different parameters through a set of algorithmic strategies, and applies different weights to the information through the attention mechanism after information fusion to obtain refined representation information. Finally, a reinforcement fusion scheme embedded in a $V^{-}$-shape generation module is used to combine the hierarchical representations to generate the missing MRI parameter. Results showed that our IMPORTANT-Net is capable of generating missing MRI parameters and outperforms comparable state-of-the-art networks. Our code is available at https://github.com/Netherlands-Cancer-Institute/MRI_IMPORTANT_NET.
☆ A Systematic Evaluation of Backdoor Trigger Characteristics in Image Classification
Deep learning achieves outstanding results in many machine learning tasks. Nevertheless, it is vulnerable to backdoor attacks that modify the training set to embed a secret functionality in the trained model. The modified training samples have a secret property, i.e., a trigger. At inference time, the secret functionality is activated when the input contains the trigger, while the model functions correctly in other cases. While there are many known backdoor attacks (and defenses), deploying a stealthy attack is still far from trivial. Successfully creating backdoor triggers heavily depends on numerous parameters. Unfortunately, research has not yet determined which parameters contribute most to the attack performance. This paper systematically analyzes the most relevant parameters for the backdoor attacks, i.e., trigger size, position, color, and poisoning rate. Using transfer learning, which is very common in computer vision, we evaluate the attack on numerous state-of-the-art models (ResNet, VGG, AlexNet, and GoogLeNet) and datasets (MNIST, CIFAR10, and TinyImageNet). Our attacks cover the majority of backdoor settings in research, providing concrete directions for future works. Our code is publicly available to facilitate the reproducibility of our results.
☆ Rethinking Semi-Supervised Medical Image Segmentation: A Variance-Reduction Perspective
For medical image segmentation, contrastive learning is the dominant practice to improve the quality of visual representations by contrasting semantically similar and dissimilar pairs of samples. This is enabled by the observation that without accessing ground truth label, negative examples with truly dissimilar anatomical features, if sampled, can significantly improve the performance. In reality, however, these samples may come from similar anatomical features and the models may struggle to distinguish the minority tail-class samples, making the tail classes more prone to misclassification, both of which typically lead to model collapse. In this paper, we propose ARCO, a semi-supervised contrastive learning (CL) framework with stratified group sampling theory in medical image segmentation. In particular, we first propose building ARCO through the concept of variance-reduced estimation, and show that certain variance-reduction techniques are particularly beneficial in medical image segmentation tasks with extremely limited labels. Furthermore, we theoretically prove these sampling techniques are universal in variance reduction. Finally, we experimentally validate our approaches on three benchmark datasets with different label settings, and our methods consistently outperform state-of-the-art semi- and fully-supervised methods. Additionally, we augment the CL frameworks with these sampling techniques and demonstrate significant gains over previous methods. We believe our work is an important step towards semi-supervised medical image segmentation by quantifying the limitation of current self-supervision objectives for accomplishing medical image analysis tasks.
☆ TEXTure: Text-Guided Texturing of 3D Shapes
In this paper, we present TEXTure, a novel method for text-guided generation, editing, and transfer of textures for 3D shapes. Leveraging a pretrained depth-to-image diffusion model, TEXTure applies an iterative scheme that paints a 3D model from different viewpoints. Yet, while depth-to-image models can create plausible textures from a single viewpoint, the stochastic nature of the generation process can cause many inconsistencies when texturing an entire 3D object. To tackle these problems, we dynamically define a trimap partitioning of the rendered image into three progression states, and present a novel elaborated diffusion sampling process that uses this trimap representation to generate seamless textures from different views. We then show that one can transfer the generated texture maps to new 3D geometries without requiring explicit surface-to-surface mapping, as well as extract semantic textures from a set of images without requiring any explicit reconstruction. Finally, we show that TEXTure can be used to not only generate new textures but also edit and refine existing textures using either a text prompt or user-provided scribbles. We demonstrate that our TEXTuring method excels at generating, transferring, and editing textures through extensive evaluation, and further close the gap between 2D image generation and 3D texturing.
comment: Project page available at https://texturepaper.github.io/TEXTurePaper/
☆ Crucial Semantic Classifier-based Adversarial Learning for Unsupervised Domain Adaptation
Unsupervised Domain Adaptation (UDA), which aims to explore the transferrable features from a well-labeled source domain to a related unlabeled target domain, has been widely progressed. Nevertheless, as one of the mainstream, existing adversarial-based methods neglect to filter the irrelevant semantic knowledge, hindering adaptation performance improvement. Besides, they require an additional domain discriminator that strives extractor to generate confused representations, but discrete designing may cause model collapse. To tackle the above issues, we propose Crucial Semantic Classifier-based Adversarial Learning (CSCAL), which pays more attention to crucial semantic knowledge transferring and leverages the classifier to implicitly play the role of domain discriminator without extra network designing. Specifically, in intra-class-wise alignment, a Paired-Level Discrepancy (PLD) is designed to transfer crucial semantic knowledge. Additionally, based on classifier predictions, a Nuclear Norm-based Discrepancy (NND) is formed that considers inter-class-wise information and improves the adaptation performance. Moreover, CSCAL can be effortlessly merged into different UDA methods as a regularizer and dramatically promote their performance.
☆ Show me your NFT and I tell you how it will perform: Multimodal representation learning for NFT selling price prediction
Non-Fungible Tokens (NFTs) represent deeds of ownership, based on blockchain technologies and smart contracts, of unique crypto assets on digital art forms (e.g., artworks or collectibles). In the spotlight after skyrocketing in 2021, NFTs have attracted the attention of crypto enthusiasts and investors intent on placing promising investments in this profitable market. However, the NFT financial performance prediction has not been widely explored to date. In this work, we address the above problem based on the hypothesis that NFT images and their textual descriptions are essential proxies to predict the NFT selling prices. To this purpose, we propose MERLIN, a novel multimodal deep learning framework designed to train Transformer-based language and visual models, along with graph neural network models, on collections of NFTs' images and texts. A key aspect in MERLIN is its independence on financial features, as it exploits only the primary data a user interested in NFT trading would like to deal with, i.e., NFT images and textual descriptions. By learning dense representations of such data, a price-category classification task is performed by MERLIN models, which can also be tuned according to user preferences in the inference phase to mimic different risk-return investment profiles. Experimental evaluation on a publicly available dataset has shown that MERLIN models achieve significant performances according to several financial assessment criteria, fostering profitable investments, and also beating baseline machine-learning classifiers based on financial features.
comment: Accepted paper at The ACM Web Conference 2023, April 30--May 04, 2023, Austin, Texas, USA
☆ CVTNet: A Cross-View Transformer Network for Place Recognition Using LiDAR Data
LiDAR-based place recognition (LPR) is one of the most crucial components of autonomous vehicles to identify previously visited places in GPS-denied environments. Most existing LPR methods use mundane representations of the input point cloud without considering different views, which may not fully exploit the information from LiDAR sensors. In this paper, we propose a cross-view transformer-based network, dubbed CVTNet, to fuse the range image views (RIVs) and bird's eye views (BEVs) generated from the LiDAR data. It extracts correlations within the views themselves using intra-transformers and between the two different views using inter-transformers. Based on that, our proposed CVTNet generates a yaw-angle-invariant global descriptor for each laser scan end-to-end online and retrieves previously seen places by descriptor matching between the current query scan and the pre-built database. We evaluate our approach on three datasets collected with different sensor setups and environmental conditions. The experimental results show that our method outperforms the state-of-the-art LPR methods with strong robustness to viewpoint changes and long-time spans. Furthermore, our approach has a good real-time performance that can run faster than the typical LiDAR frame rate. The implementation of our method is released as open source at: https://github.com/BIT-MJY/CVTNet.
☆ The Learnable Typewriter: A Generative Approach to Text Line Analysis
We present a generative document-specific approach to character analysis and recognition in text lines. Our main idea is to build on unsupervised multi-object segmentation methods and in particular those that reconstruct images based on a limited amount of visual elements, called sprites. Our approach can learn a large number of different characters and leverage line-level annotations when available. Our contribution is twofold. First, we provide the first adaptation and evaluation of a deep unsupervised multi-object segmentation approach for text line analysis. Since these methods have mainly been evaluated on synthetic data in a completely unsupervised setting, demonstrating that they can be adapted and quantitatively evaluated on real text images and that they can be trained using weak supervision are significant progresses. Second, we demonstrate the potential of our method for new applications, more specifically in the field of paleography, which studies the history and variations of handwriting, and for cipher analysis. We evaluate our approach on three very different datasets: a printed volume of the Google1000 dataset, the Copiale cipher and historical handwritten charters from the 12th and early 13th century.
☆ From slides (through tiles) to pixels: an explainability framework for weakly supervised models in pre-clinical pathology
In pre-clinical pathology, there is a paradox between the abundance of raw data (whole slide images from many organs of many individual animals) and the lack of pixel-level slide annotations done by pathologists. Due to time constraints and requirements from regulatory authorities, diagnoses are instead stored as slide labels. Weakly supervised training is designed to take advantage of those data, and the trained models can be used by pathologists to rank slides by their probability of containing a given lesion of interest. In this work, we propose a novel contextualized eXplainable AI (XAI) framework and its application to deep learning models trained on Whole Slide Images (WSIs) in Digital Pathology. Specifically, we apply our methods to a multi-instance-learning (MIL) model, which is trained solely on slide-level labels, without the need for pixel-level annotations. We validate quantitatively our methods by quantifying the agreements of our explanations' heatmaps with pathologists' annotations, as well as with predictions from a segmentation model trained on such annotations. We demonstrate the stability of the explanations with respect to input shifts, and the fidelity with respect to increased model performance. We quantitatively evaluate the correlation between available pixel-wise annotations and explainability heatmaps. We show that the explanations on important tiles of the whole slide correlate with tissue changes between healthy regions and lesions, but do not exactly behave like a human annotator. This result is coherent with the model training strategy.
comment: 18 pages, 9 figures
☆ ShadowFormer: Global Context Helps Image Shadow Removal AAAI2023
Recent deep learning methods have achieved promising results in image shadow removal. However, most of the existing approaches focus on working locally within shadow and non-shadow regions, resulting in severe artifacts around the shadow boundaries as well as inconsistent illumination between shadow and non-shadow regions. It is still challenging for the deep shadow removal model to exploit the global contextual correlation between shadow and non-shadow regions. In this work, we first propose a Retinex-based shadow model, from which we derive a novel transformer-based network, dubbed ShandowFormer, to exploit non-shadow regions to help shadow region restoration. A multi-scale channel attention framework is employed to hierarchically capture the global information. Based on that, we propose a Shadow-Interaction Module (SIM) with Shadow-Interaction Attention (SIA) in the bottleneck stage to effectively model the context correlation between shadow and non-shadow regions. We conduct extensive experiments on three popular public datasets, including ISTD, ISTD+, and SRD, to evaluate the proposed method. Our method achieves state-of-the-art performance by using up to 150X fewer model parameters.
comment: Accepted by AAAI2023
☆ A statistically constrained internal method for single image super-resolution
Deep learning based methods for single-image super-resolution (SR) have drawn a lot of attention lately. In particular, various papers have shown that the learning stage can be performed on a single image, resulting in the so-called internal approaches. The SinGAN method is one of these contributions, where the distribution of image patches is learnt on the image at hand and propagated at finer scales. Now, there are situations where some statistical a priori can be assumed for the final image. In particular, many natural phenomena yield images having power law Fourier spectrum, such as clouds and other texture images. In this work, we show how such a priori information can be integrated into an internal super-resolution approach, by constraining the learned up-sampling procedure of SinGAN. We consider various types of constraints, related to the Fourier power spectrum, the color histograms and the consistency of the upsampling scheme. We demonstrate on various experiments that these constraints are indeed satisfied, but also that some perceptual quality measures can be improved by the proposed approach.
Blockwise Self-Supervised Learning at Scale
Current state-of-the-art deep networks are all powered by backpropagation. In this paper, we explore alternatives to full backpropagation in the form of blockwise learning rules, leveraging the latest developments in self-supervised learning. We show that a blockwise pretraining procedure consisting of training independently the 4 main blocks of layers of a ResNet-50 with Barlow Twins' loss function at each block performs almost as well as end-to-end backpropagation on ImageNet: a linear probe trained on top of our blockwise pretrained model obtains a top-1 classification accuracy of 70.48%, only 1.1% below the accuracy of an end-to-end pretrained network (71.57% accuracy). We perform extensive experiments to understand the impact of different components within our method and explore a variety of adaptations of self-supervised learning to the blockwise paradigm, building an exhaustive understanding of the critical avenues for scaling local learning rules to large networks, with implications ranging from hardware design to neuroscience.
☆ Cluster-CAM: Cluster-Weighted Visual Interpretation of CNNs' Decision in Image Classification
Despite the tremendous success of convolutional neural networks (CNNs) in computer vision, the mechanism of CNNs still lacks clear interpretation. Currently, class activation mapping (CAM), a famous visualization technique to interpret CNN's decision, has drawn increasing attention. Gradient-based CAMs are efficient while the performance is heavily affected by gradient vanishing and exploding. In contrast, gradient-free CAMs can avoid computing gradients to produce more understandable results. However, existing gradient-free CAMs are quite time-consuming because hundreds of forward interference per image are required. In this paper, we proposed Cluster-CAM, an effective and efficient gradient-free CNN interpretation algorithm. Cluster-CAM can significantly reduce the times of forward propagation by splitting the feature maps into clusters in an unsupervised manner. Furthermore, we propose an artful strategy to forge a cognition-base map and cognition-scissors from clustered feature maps. The final salience heatmap will be computed by merging the above cognition maps. Qualitative results conspicuously show that Cluster-CAM can produce heatmaps where the highlighted regions match the human's cognition more precisely than existing CAMs. The quantitative evaluation further demonstrates the superiority of Cluster-CAM in both effectiveness and efficiency.
comment: 10 pages
☆ Private, fair and accurate: Training large-scale, privacy-preserving AI models in radiology
Artificial intelligence (AI) models are increasingly used in the medical domain. However, as medical data is highly sensitive, special precautions to ensure the protection of said data are required. The gold standard for privacy preservation is the introduction of differential privacy (DP) to model training. However, prior work has shown that DP has negative implications on model accuracy and fairness. Therefore, the purpose of this study is to demonstrate that the privacy-preserving training of AI models for chest radiograph diagnosis is possible with high accuracy and fairness compared to non-private training. N=193,311 high quality clinical chest radiographs were retrospectively collected and manually labeled by experienced radiologists, who assigned one or more of the following diagnoses: cardiomegaly, congestion, pleural effusion, pneumonic infiltration and atelectasis, to each side (where applicable). The non-private AI models were compared with privacy-preserving (DP) models with respect to privacy-utility trade-offs (measured as area under the receiver-operator-characteristic curve (AUROC)), and privacy-fairness trade-offs (measured as Pearson-R or Statistical Parity Difference). The non-private AI model achieved an average AUROC score of 0.90 over all labels, whereas the DP AI model with a privacy budget of epsilon=7.89 resulted in an AUROC of 0.87, i.e., a mere 2.6% performance decrease compared to non-private training. The privacy-preserving training of diagnostic AI models can achieve high performance with a small penalty on model accuracy and does not amplify discrimination against age, sex or co-morbidity. We thus encourage practitioners to integrate state-of-the-art privacy-preserving techniques into medical AI model development.
comment: 3 tables, 5 figures, 11 supplementary materials
☆ A geometrically aware auto-encoder for multi-texture synthesis
We propose an auto-encoder architecture for multi-texture synthesis. The approach relies on both a compact encoder accounting for second order neural statistics and a generator incorporating adaptive periodic content. Images are embedded in a compact and geometrically consistent latent space, where the texture representation and its spatial organisation are disentangled. Texture synthesis and interpolation tasks can be performed directly from these latent codes. Our experiments demonstrate that our model outperforms state-of-the-art feed-forward methods in terms of visual quality and various texture related metrics.
☆ CFFT-GAN: Cross-domain Feature Fusion Transformer for Exemplar-based Image Translation AAAI2023
Exemplar-based image translation refers to the task of generating images with the desired style, while conditioning on certain input image. Most of the current methods learn the correspondence between two input domains and lack the mining of information within the domains. In this paper, we propose a more general learning approach by considering two domain features as a whole and learning both inter-domain correspondence and intra-domain potential information interactions. Specifically, we propose a Cross-domain Feature Fusion Transformer (CFFT) to learn inter- and intra-domain feature fusion. Based on CFFT, the proposed CFFT-GAN works well on exemplar-based image translation. Moreover, CFFT-GAN is able to decouple and fuse features from multiple domains by cascading CFFT modules. We conduct rich quantitative and qualitative experiments on several image translation tasks, and the results demonstrate the superiority of our approach compared to state-of-the-art methods. Ablation studies show the importance of our proposed CFFT. Application experimental results reflect the potential of our method.
comment: Accepted by AAAI2023
☆ A Feature Selection Method for Driver Stress Detection Using Heart Rate Variability and Breathing Rate
Driver stress is a major cause of car accidents and death worldwide. Furthermore, persistent stress is a health problem, contributing to hypertension and other diseases of the cardiovascular system. Stress has a measurable impact on heart and breathing rates and stress levels can be inferred from such measurements. Galvanic skin response is a common test to measure the perspiration caused by both physiological and psychological stress, as well as extreme emotions. In this paper, galvanic skin response is used to estimate the ground truth stress levels. A feature selection technique based on the minimal redundancy-maximal relevance method is then applied to multiple heart rate variability and breathing rate metrics to identify a novel and optimal combination for use in detecting stress. The support vector machine algorithm with a radial basis function kernel was used along with these features to reliably predict stress. The proposed method has achieved a high level of accuracy on the target dataset.
comment: In Proceedings of the 15th International Conference on Machine Vision (ICMV), Romem Italy, 18-20 November 2022. arXiv admin note: text overlap with arXiv:2206.03222
☆ Explicit Box Detection Unifies End-to-End Multi-Person Pose Estimation ICLR 2023
This paper presents a novel end-to-end framework with Explicit box Detection for multi-person Pose estimation, called ED-Pose, where it unifies the contextual learning between human-level (global) and keypoint-level (local) information. Different from previous one-stage methods, ED-Pose re-considers this task as two explicit box detection processes with a unified representation and regression supervision. First, we introduce a human detection decoder from encoded tokens to extract global features. It can provide a good initialization for the latter keypoint detection, making the training process converge fast. Second, to bring in contextual information near keypoints, we regard pose estimation as a keypoint box detection problem to learn both box positions and contents for each keypoint. A human-to-keypoint detection decoder adopts an interactive learning strategy between human and keypoint features to further enhance global and local feature aggregation. In general, ED-Pose is conceptually simple without post-processing and dense heatmap supervision. It demonstrates its effectiveness and efficiency compared with both two-stage and one-stage methods. Notably, explicit box detection boosts the pose estimation performance by 4.5 AP on COCO and 9.9 AP on CrowdPose. For the first time, as a fully end-to-end framework with a L1 regression loss, ED-Pose surpasses heatmap-based Top-down methods under the same backbone by 1.2 AP on COCO and achieves the state-of-the-art with 76.6 AP on CrowdPose without bells and whistles. Code is available at https://github.com/IDEA-Research/ED-Pose.
comment: Accepted to ICLR 2023
☆ SegForestNet: Spatial-Partitioning-Based Aerial Image Segmentation
Aerial image analysis, specifically the semantic segmentation thereof, is the basis for applications such as automatically creating and updating maps, tracking city growth, or tracking deforestation. In true orthophotos, which are often used in these applications, many objects and regions can be approximated well by polygons. However, this fact is rarely exploited by state-of-the-art semantic segmentation models. Instead, most models allow unnecessary degrees of freedom in their predictions by allowing arbitrary region shapes. We therefore present a refinement of our deep learning model which predicts binary space partitioning trees, an efficient polygon representation. The refinements include a new feature decoder architecture and a new differentiable BSP tree renderer which both avoid vanishing gradients. Additionally, we designed a novel loss function specifically designed to improve the spatial partitioning defined by the predicted trees. Furthermore, our expanded model can predict multiple trees at once and thus can predict class-specific segmentations. Taking all modifications together, our model achieves state-of-the-art performance while using up to 60% fewer model parameters when using a small backbone model or up to 20% fewer model parameters when using a large backbone model.
comment: This work has been submitted to the IEEE for possible publication. Copyright may be transferred without notice, after which this version may no longer be accessible
☆ Semantic 3D-aware Portrait Synthesis and Manipulation Based on Compositional Neural Radiance Field AAAI2023
Recently 3D-aware GAN methods with neural radiance field have developed rapidly. However, current methods model the whole image as an overall neural radiance field, which limits the partial semantic editability of synthetic results. Since NeRF renders an image pixel by pixel, it is possible to split NeRF in the spatial dimension. We propose a Compositional Neural Radiance Field (CNeRF) for semantic 3D-aware portrait synthesis and manipulation. CNeRF divides the image by semantic regions and learns an independent neural radiance field for each region, and finally fuses them and renders the complete image. Thus we can manipulate the synthesized semantic regions independently, while fixing the other parts unchanged. Furthermore, CNeRF is also designed to decouple shape and texture within each semantic region. Compared to state-of-the-art 3D-aware GAN methods, our approach enables fine-grained semantic region manipulation, while maintaining high-quality 3D-consistent synthesis. The ablation studies show the effectiveness of the structure and loss function used by our method. In addition real image inversion and cartoon portrait 3D editing experiments demonstrate the application potential of our method.
comment: Accepted by AAAI2023
☆ Simple, Effective and General: A New Backbone for Cross-view Image Geo-localization
In this work, we aim at an important but less explored problem of a simple yet effective backbone specific for cross-view geo-localization task. Existing methods for cross-view geo-localization tasks are frequently characterized by 1) complicated methodologies, 2) GPU-consuming computations, and 3) a stringent assumption that aerial and ground images are centrally or orientation aligned. To address the above three challenges for cross-view image matching, we propose a new backbone network, named Simple Attention-based Image Geo-localization network (SAIG). The proposed SAIG effectively represents long-range interactions among patches as well as cross-view correspondence with multi-head self-attention layers. The "narrow-deep" architecture of our SAIG improves the feature richness without degradation in performance, while its shallow and effective convolutional stem preserves the locality, eliminating the loss of patchify boundary information. Our SAIG achieves state-of-the-art results on cross-view geo-localization, while being far simpler than previous works. Furthermore, with only 15.9% of the model parameters and half of the output dimension compared to the state-of-the-art, the SAIG adapts well across multiple cross-view datasets without employing any well-designed feature aggregation modules or feature alignment algorithms. In addition, our SAIG attains competitive scores on image retrieval benchmarks, further demonstrating its generalizability. As a backbone network, our SAIG is both easy to follow and computationally lightweight, which is meaningful in practical scenario. Moreover, we propose a simple Spatial-Mixed feature aggregation moDule (SMD) that can mix and project spatial information into a low-dimensional space to generate feature descriptors... (The code is available at https://github.com/yanghongji2007/SAIG)
comment: Under Review
☆ Robust Camera Pose Refinement for Multi-Resolution Hash Encoding
Multi-resolution hash encoding has recently been proposed to reduce the computational cost of neural renderings, such as NeRF. This method requires accurate camera poses for the neural renderings of given scenes. However, contrary to previous methods jointly optimizing camera poses and 3D scenes, the naive gradient-based camera pose refinement method using multi-resolution hash encoding severely deteriorates performance. We propose a joint optimization algorithm to calibrate the camera pose and learn a geometric representation using efficient multi-resolution hash encoding. Showing that the oscillating gradient flows of hash encoding interfere with the registration of camera poses, our method addresses the issue by utilizing smooth interpolation weighting to stabilize the gradient oscillation for the ray samplings across hash grids. Moreover, the curriculum training procedure helps to learn the level-wise hash encoding, further increasing the pose refinement. Experiments on the novel-view synthesis datasets validate that our learning frameworks achieve state-of-the-art performance and rapid convergence of neural rendering, even when initial camera poses are unknown.
☆ Bridging the Emotional Semantic Gap via Multimodal Relevance Estimation
Human beings have rich ways of emotional expressions, including facial action, voice, and natural languages. Due to the diversity and complexity of different individuals, the emotions expressed by various modalities may be semantically irrelevant. Directly fusing information from different modalities may inevitably make the model subject to the noise from semantically irrelevant modalities. To tackle this problem, we propose a multimodal relevance estimation network to capture the relevant semantics among modalities in multimodal emotions. Specifically, we take advantage of an attention mechanism to reflect the semantic relevance weights of each modality. Moreover, we propose a relevant semantic estimation loss to weakly supervise the semantics of each modality. Furthermore, we make use of contrastive learning to optimize the similarity of category-level modality-relevant semantics across different modalities in feature space, thereby bridging the semantic gap between heterogeneous modalities. In order to better reflect the emotional state in the real interactive scenarios and perform the semantic relevance analysis, we collect a single-label discrete multimodal emotion dataset named SDME, which enables researchers to conduct multimodal semantic relevance research with large category bias. Experiments on continuous and discrete emotion datasets show that our model can effectively capture the relevant semantics, especially for the large deviations in modal semantics. The code and SDME dataset will be publicly available.
☆ Contrastive Learning with Consistent Representations
Contrastive learning demonstrates great promise for representation learning. Data augmentations play a critical role in contrastive learning by providing informative views of the data without needing the labels. However, the performance of the existing works heavily relies on the quality of the employed data augmentation (DA) functions, which are typically hand picked from a restricted set of choices. While exploiting a diverse set of data augmentations is appealing, the intricacies of DAs and representation learning may lead to performance degradation. To address this challenge and allow for a systemic use of large numbers of data augmentations, this paper proposes Contrastive Learning with Consistent Representations (CoCor). At the core of CoCor is a new consistency measure, DA consistency, which dictates the mapping of augmented input data to the representation space such that these instances are mapped to optimal locations in a way consistent to the intensity of the DA applied. Furthermore, a data-driven approach is proposed to learn the optimal mapping locations as a function of DA while maintaining a desired monotonic property with respect to DA intensity. The proposed techniques give rise to a semi-supervised learning framework based on bi-level optimization, achieving new state-of-the-art results for image recognition.
☆ DEVICE: DEpth and VIsual ConcEpts Aware Transformer for TextCaps
Text-based image captioning is an important but under-explored task, aiming to generate descriptions containing visual objects and scene text. Recent studies have made encouraging progress, but they are still suffering from a lack of overall understanding of scenes and generating inaccurate captions. One possible reason is that current studies mainly focus on constructing the plane-level geometric relationship of scene text without depth information. This leads to insufficient scene text relational reasoning so that models may describe scene text inaccurately. The other possible reason is that existing methods fail to generate fine-grained descriptions of some visual objects. In addition, they may ignore essential visual objects, leading to the scene text belonging to these ignored objects not being utilized. To address the above issues, we propose a DEpth and VIsual ConcEpts Aware Transformer (DEVICE) for TextCaps. Concretely, to construct three-dimensional geometric relations, we introduce depth information and propose a depth-enhanced feature updating module to ameliorate OCR token features. To generate more precise and comprehensive captions, we introduce semantic features of detected visual object concepts as auxiliary information. Our DEVICE is capable of generalizing scenes more comprehensively and boosting the accuracy of described visual entities. Sufficient experiments demonstrate the effectiveness of our proposed DEVICE, which outperforms state-of-the-art models on the TextCaps test set. Our code will be publicly available.
comment: 9pages, 5 figures
☆ INV: Towards Streaming Incremental Neural Videos
Recent works in spatiotemporal radiance fields can produce photorealistic free-viewpoint videos. However, they are inherently unsuitable for interactive streaming scenarios (e.g. video conferencing, telepresence) because have an inevitable lag even if the training is instantaneous. This is because these approaches consume videos and thus have to buffer chunks of frames (often seconds) before processing. In this work, we take a step towards interactive streaming via a frame-by-frame approach naturally free of lag. Conventional wisdom believes that per-frame NeRFs are impractical due to prohibitive training costs and storage. We break this belief by introducing Incremental Neural Videos (INV), a per-frame NeRF that is efficiently trained and streamable. We designed INV based on two insights: (1) Our main finding is that MLPs naturally partition themselves into Structure and Color Layers, which store structural and color/texture information respectively. (2) We leverage this property to retain and improve upon knowledge from previous frames, thus amortizing training across frames and reducing redundant learning. As a result, with negligible changes to NeRF, INV can achieve good qualities (>28.6db) in 8min/frame. It can also outperform prior SOTA in 19% less training time. Additionally, our Temporal Weight Compression reduces the per-frame size to 0.3MB/frame (6.6% of NeRF). More importantly, INV is free from buffer lag and is naturally fit for streaming. While this work does not achieve real-time training, it shows that incremental approaches like INV present new possibilities in interactive 3D streaming. Moreover, our discovery of natural information partition leads to a better understanding and manipulation of MLPs. Code and dataset will be released soon.
☆ Example-Based Explainable AI and its Application for Remote Sensing Image Classification
We present a method of explainable artificial intelligence (XAI), "What I Know (WIK)", to provide additional information to verify the reliability of a deep learning model by showing an example of an instance in a training dataset that is similar to the input data to be inferred and demonstrate it in a remote sensing image classification task. One of the expected roles of XAI methods is verifying whether inferences of a trained machine learning model are valid for an application, and it is an important factor that what datasets are used for training the model as well as the model architecture. Our data-centric approach can help determine whether the training dataset is sufficient for each inference by checking the selected example data. If the selected example looks similar to the input data, we can confirm that the model was not trained on a dataset with a feature distribution far from the feature of the input data. With this method, the criteria for selecting an example are not merely data similarity with the input data but also data similarity in the context of the model task. Using a remote sensing image dataset from the Sentinel-2 satellite, the concept was successfully demonstrated with reasonably selected examples. This method can be applied to various machine-learning tasks, including classification and regression.
comment: 10 pages, 4 figures, accepted for publication in International Journal of Applied Earth Observation and Geoinformation
☆ Class Overwhelms: Mutual Conditional Blended-Target Domain Adaptation
Current methods of blended targets domain adaptation (BTDA) usually infer or consider domain label information but underemphasize hybrid categorical feature structures of targets, which yields limited performance, especially under the label distribution shift. We demonstrate that domain labels are not directly necessary for BTDA if categorical distributions of various domains are sufficiently aligned even facing the imbalance of domains and the label distribution shift of classes. However, we observe that the cluster assumption in BTDA does not comprehensively hold. The hybrid categorical feature space hinders the modeling of categorical distributions and the generation of reliable pseudo labels for categorical alignment. To address these, we propose a categorical domain discriminator guided by uncertainty to explicitly model and directly align categorical distributions $P(Z|Y)$. Simultaneously, we utilize the low-level features to augment the single source features with diverse target styles to rectify the biased classifier $P(Y|Z)$ among diverse targets. Such a mutual conditional alignment of $P(Z|Y)$ and $P(Y|Z)$ forms a mutual reinforced mechanism. Our approach outperforms the state-of-the-art in BTDA even compared with methods utilizing domain labels, especially under the label distribution shift, and in single target DA on DomainNet.
☆ Spectral Aware Softmax for Visible-Infrared Person Re-Identification
Visible-infrared person re-identification (VI-ReID) aims to match specific pedestrian images from different modalities. Although suffering an extra modality discrepancy, existing methods still follow the softmax loss training paradigm, which is widely used in single-modality classification tasks. The softmax loss lacks an explicit penalty for the apparent modality gap, which adversely limits the performance upper bound of the VI-ReID task. In this paper, we propose the spectral-aware softmax (SA-Softmax) loss, which can fully explore the embedding space with the modality information and has clear interpretability. Specifically, SA-Softmax loss utilizes an asynchronous optimization strategy based on the modality prototype instead of the synchronous optimization based on the identity prototype in the original softmax loss. To encourage a high overlapping between two modalities, SA-Softmax optimizes each sample by the prototype from another spectrum. Based on the observation and analysis of SA-Softmax, we modify the SA-Softmax with the Feature Mask and Absolute-Similarity Term to alleviate the ambiguous optimization during model training. Extensive experimental evaluations conducted on RegDB and SYSU-MM01 demonstrate the superior performance of the SA-Softmax over the state-of-the-art methods in such a cross-modality condition.
☆ Revisiting Long-tailed Image Classification: Survey and Benchmarks with New Evaluation Metrics
Recently, long-tailed image classification harvests lots of research attention, since the data distribution is long-tailed in many real-world situations. Piles of algorithms are devised to address the data imbalance problem by biasing the training process towards less frequent classes. However, they usually evaluate the performance on a balanced testing set or multiple independent testing sets having distinct distributions with the training data. Considering the testing data may have arbitrary distributions, existing evaluation strategies are unable to reflect the actual classification performance objectively. We set up novel evaluation benchmarks based on a series of testing sets with evolving distributions. A corpus of metrics are designed for measuring the accuracy, robustness, and bounds of algorithms for learning with long-tailed distribution. Based on our benchmarks, we re-evaluate the performance of existing methods on CIFAR10 and CIFAR100 datasets, which is valuable for guiding the selection of data rebalancing techniques. We also revisit existing methods and categorize them into four types including data balancing, feature balancing, loss balancing, and prediction balancing, according the focused procedure during the training pipeline.
☆ Deep Learning (DL)-based Automatic Segmentation of the Internal Pudendal Artery (IPA) for Reduction of Erectile Dysfunction in Definitive Radiotherapy of Localized Prostate Cancer
Background and purpose: Radiation-induced erectile dysfunction (RiED) is commonly seen in prostate cancer patients. Clinical trials have been developed in multiple institutions to investigate whether dose-sparing to the internal-pudendal-arteries (IPA) will improve retention of sexual potency. The IPA is usually not considered a conventional organ-at-risk (OAR) due to segmentation difficulty. In this work, we propose a deep learning (DL)-based auto-segmentation model for the IPA that utilizes CT and MRI or CT alone as the input image modality to accommodate variation in clinical practice. Materials and methods: 86 patients with CT and MRI images and noisy IPA labels were recruited in this study. We split the data into 42/14/30 for model training, testing, and a clinical observer study, respectively. There were three major innovations in this model: 1) we designed an architecture with squeeze-and-excite blocks and modality attention for effective feature extraction and production of accurate segmentation, 2) a novel loss function was used for training the model effectively with noisy labels, and 3) modality dropout strategy was used for making the model capable of segmentation in the absence of MRI. Results: The DSC, ASD, and HD95 values for the test dataset were 62.2%, 2.54mm, and 7mm, respectively. AI segmented contours were dosimetrically equivalent to the expert physician's contours. The observer study showed that expert physicians' scored AI contours (mean=3.7) higher than inexperienced physicians' contours (mean=3.1). When inexperienced physicians started with AI contours, the score improved to 3.7. Conclusion: The proposed model achieved good quality IPA contours to improve uniformity of segmentation and to facilitate introduction of standardized IPA segmentation into clinical trials and practice.
♻ ☆ One-shot domain adaptation in video-based assessment of surgical skills
Deep Learning (DL) has achieved automatic and objective assessment of surgical skills. However, DL models are data-hungry and restricted to their training domain. This prevents them from transitioning to new tasks where data is limited. Hence, domain adaptation is crucial to implement DL in real life. Here, we propose a meta-learning model, A-VBANet, that can deliver domain-agnostic surgical skill classification via one-shot learning. We develop the A-VBANet on five laparoscopic and robotic surgical simulators. Additionally, we test it on operating room (OR) videos of laparoscopic cholecystectomy. Our model successfully adapts with accuracies up to 99.5% in one-shot and 99.9% in few-shot settings for simulated tasks and 89.7% for laparoscopic cholecystectomy. For the first time, we provide a domain-agnostic procedure for video-based assessment of surgical skills. A significant implication of this approach is that it allows the use of data from surgical simulators to assess performance in the operating room.
comment: 12 pages (+9 pages of Supplementary Materials), 4 figures (+2 Supplementary Figures), 2 tables (+5 Supplementary Tables)
♻ ☆ Open-Set Multi-Source Multi-Target Domain Adaptation NeurIPS 2021
Single-Source Single-Target Domain Adaptation (1S1T) aims to bridge the gap between a labelled source domain and an unlabelled target domain. Despite 1S1T being a well-researched topic, they are typically not deployed to the real world. Methods like Multi-Source Domain Adaptation and Multi-Target Domain Adaptation have evolved to model real-world problems but still do not generalise well. The fact that most of these methods assume a common label-set between source and target is very restrictive. Recent Open-Set Domain Adaptation methods handle unknown target labels but fail to generalise in multiple domains. To overcome these difficulties, first, we propose a novel generic domain adaptation (DA) setting named Open-Set Multi-Source Multi-Target Domain Adaptation (OS-nSmT), with n and m being number of source and target domains respectively. Next, we propose a graph attention based framework named DEGAA which can capture information from multiple source and target domains without knowing the exact label-set of the target. We argue that our method, though offered for multiple sources and multiple targets, can also be agnostic to various other DA settings. To check the robustness and versatility of DEGAA, we put forward ample experiments and ablation studies.
comment: Accepted in NeurIPS 2021 Workshop on Pre-registration in Machine Learning
♻ ☆ Faster Attention Is What You Need: A Fast Self-Attention Neural Network Backbone Architecture for the Edge via Double-Condensing Attention Condensers
With the growing adoption of deep learning for on-device TinyML applications, there has been an ever-increasing demand for efficient neural network backbones optimized for the edge. Recently, the introduction of attention condenser networks have resulted in low-footprint, highly-efficient, self-attention neural networks that strike a strong balance between accuracy and speed. In this study, we introduce a faster attention condenser design called double-condensing attention condensers that allow for highly condensed feature embeddings. We further employ a machine-driven design exploration strategy that imposes design constraints based on best practices for greater efficiency and robustness to produce the macro-micro architecture constructs of the backbone. The resulting backbone (which we name AttendNeXt) achieves significantly higher inference throughput on an embedded ARM processor when compared to several other state-of-the-art efficient backbones (>10x faster than FB-Net C at higher accuracy and speed and >10x faster than MobileOne-S1 at smaller size) while having a small model size (>1.37x smaller than MobileNetv3-L at higher accuracy and speed) and strong accuracy (1.1% higher top-1 accuracy than MobileViT XS on ImageNet at higher speed). These promising results demonstrate that exploring different efficient architecture designs and self-attention mechanisms can lead to interesting new building blocks for TinyML applications.
♻ ☆ Accurate Image Restoration with Attention Retractable Transformer ICLR 2023
Recently, Transformer-based image restoration networks have achieved promising improvements over convolutional neural networks due to parameter-independent global interactions. To lower computational cost, existing works generally limit self-attention computation within non-overlapping windows. However, each group of tokens are always from a dense area of the image. This is considered as a dense attention strategy since the interactions of tokens are restrained in dense regions. Obviously, this strategy could result in restricted receptive fields. To address this issue, we propose Attention Retractable Transformer (ART) for image restoration, which presents both dense and sparse attention modules in the network. The sparse attention module allows tokens from sparse areas to interact and thus provides a wider receptive field. Furthermore, the alternating application of dense and sparse attention modules greatly enhances representation ability of Transformer while providing retractable attention on the input image.We conduct extensive experiments on image super-resolution, denoising, and JPEG compression artifact reduction tasks. Experimental results validate that our proposed ART outperforms state-of-the-art methods on various benchmark datasets both quantitatively and visually. We also provide code and models at https://github.com/gladzhang/ART.
comment: Accepted to ICLR 2023. Code and models are available at https://github.com/gladzhang/ART
♻ ☆ Wise-IoU: Bounding Box Regression Loss with Dynamic Focusing Mechanism
The loss function for bounding box regression (BBR) is essential to object detection. Its good definition will bring significant performance improvement to the model. Most existing works assume that the examples in the training data are high-quality and focus on strengthening the fitting ability of BBR loss. If we blindly strengthen BBR on low-quality examples, it will jeopardize localization performance. Focal-EIoU v1 was proposed to solve this problem, but due to its static focusing mechanism (FM), the potential of non-monotonic FM was not fully exploited. Based on this idea, we propose an IoU-based loss with a dynamic non-monotonic FM named Wise-IoU (WIoU). The dynamic non-monotonic FM uses the outlier degree instead of IoU to evaluate the quality of anchor boxes and provides a wise gradient gain allocation strategy. This strategy reduces the competitiveness of high-quality anchor boxes while also reducing the harmful gradient generated by low-quality examples. This allows WIoU to focus on ordinary-quality anchor boxes and improve the detector's overall performance. When WIoU is applied to the state-of-the-art real-time detector YOLOv7, the AP-75 on the MS-COCO dataset is improved from 53.03% to 54.50%. Code is available at https://github.com/Instinct323/wiou.
♻ ☆ Semantics-driven Attentive Few-shot Learning over Clean and Noisy Samples
Over the last couple of years few-shot learning (FSL) has attracted great attention towards minimizing the dependency on labeled training examples. An inherent difficulty in FSL is the handling of ambiguities resulting from having too few training samples per class. To tackle this fundamental challenge in FSL, we aim to train meta-learner models that can leverage prior semantic knowledge about novel classes to guide the classifier synthesis process. In particular, we propose semantically-conditioned feature attention and sample attention mechanisms that estimate the importance of representation dimensions and training instances. We also study the problem of sample noise in FSL, towards the utilization of meta-learners in more realistic and imperfect settings. Our experimental results demonstrate the effectiveness of the proposed semantic FSL model with and without sample noise.
comment: 25 pages, 4 figures
♻ ☆ PDEBENCH: An Extensive Benchmark for Scientific Machine Learning NeurIPS 2022
Machine learning-based modeling of physical systems has experienced increased interest in recent years. Despite some impressive progress, there is still a lack of benchmarks for Scientific ML that are easy to use but still challenging and representative of a wide range of problems. We introduce PDEBench, a benchmark suite of time-dependent simulation tasks based on Partial Differential Equations (PDEs). PDEBench comprises both code and data to benchmark the performance of novel machine learning models against both classical numerical simulations and machine learning baselines. Our proposed set of benchmark problems contribute the following unique features: (1) A much wider range of PDEs compared to existing benchmarks, ranging from relatively common examples to more realistic and difficult problems; (2) much larger ready-to-use datasets compared to prior work, comprising multiple simulation runs across a larger number of initial and boundary conditions and PDE parameters; (3) more extensible source codes with user-friendly APIs for data generation and baseline results with popular machine learning models (FNO, U-Net, PINN, Gradient-Based Inverse Method). PDEBench allows researchers to extend the benchmark freely for their own purposes using a standardized API and to compare the performance of new models to existing baseline methods. We also propose new evaluation metrics with the aim to provide a more holistic understanding of learning methods in the context of Scientific ML. With those metrics we identify tasks which are challenging for recent ML methods and propose these tasks as future challenges for the community. The code is available at https://github.com/pdebench/PDEBench.
comment: 16 pages (main body) + 34 pages (supplemental material), accepted for publication in NeurIPS 2022 Track Datasets and Benchmarks
♻ ☆ High-resolution Iterative Feedback Network for Camouflaged Object Detection
Spotting camouflaged objects that are visually assimilated into the background is tricky for both object detection algorithms and humans who are usually confused or cheated by the perfectly intrinsic similarities between the foreground objects and the background surroundings. To tackle this challenge, we aim to extract the high-resolution texture details to avoid the detail degradation that causes blurred vision in edges and boundaries. We introduce a novel HitNet to refine the low-resolution representations by high-resolution features in an iterative feedback manner, essentially a global loop-based connection among the multi-scale resolutions. In addition, an iterative feedback loss is proposed to impose more constraints on each feedback connection. Extensive experiments on four challenging datasets demonstrate that our \ourmodel~breaks the performance bottleneck and achieves significant improvements compared with 29 state-of-the-art methods. To address the data scarcity in camouflaged scenarios, we provide an application example by employing cross-domain learning to extract the features that can reflect the camouflaged object properties and embed the features into salient objects, thereby generating more camouflaged training samples from the diverse salient object datasets The code will be available at https://github.com/HUuxiaobin/HitNet.
♻ ☆ CrossDTR: Cross-view and Depth-guided Transformers for 3D Object Detection ICRA
To achieve accurate 3D object detection at a low cost for autonomous driving, many multi-camera methods have been proposed and solved the occlusion problem of monocular approaches. However, due to the lack of accurate estimated depth, existing multi-camera methods often generate multiple bounding boxes along a ray of depth direction for difficult small objects such as pedestrians, resulting in an extremely low recall. Furthermore, directly applying depth prediction modules to existing multi-camera methods, generally composed of large network architectures, cannot meet the real-time requirements of self-driving applications. To address these issues, we propose Cross-view and Depth-guided Transformers for 3D Object Detection, CrossDTR. First, our lightweight depth predictor is designed to produce precise object-wise sparse depth maps and low-dimensional depth embeddings without extra depth datasets during supervision. Second, a cross-view depth-guided transformer is developed to fuse the depth embeddings as well as image features from cameras of different views and generate 3D bounding boxes. Extensive experiments demonstrated that our method hugely surpassed existing multi-camera methods by 10 percent in pedestrian detection and about 3 percent in overall mAP and NDS metrics. Also, computational analyses showed that our method is 5 times faster than prior approaches. Our codes will be made publicly available at https://github.com/sty61010/CrossDTR.
comment: Accepted by IEEE International Conference on Robotics and Automation (ICRA) 2023. The code is available at https://github.com/sty61010/CrossDTR
♻ ☆ Searching for the Essence of Adversarial Perturbations
Neural networks have demonstrated state-of-the-art performance in various machine learning fields. However, the introduction of malicious perturbations in input data, known as adversarial examples, has been shown to deceive neural network predictions. This poses potential risks for real-world applications such as autonomous driving and text identification. In order to mitigate these risks, a comprehensive understanding of the mechanisms underlying adversarial examples is essential. In this study, we demonstrate that adversarial perturbations contain human-recognizable information, which is the key conspirator responsible for a neural network's incorrect prediction, in contrast to the widely held belief that human-unidentifiable characteristics play a critical role in fooling a network. This concept of human-recognizable characteristics enables us to explain key features of adversarial perturbations, including their existence, transferability among different neural networks, and increased interpretability for adversarial training. We also uncover two unique properties of adversarial perturbations that deceive neural networks: masking and generation. Additionally, a special class, the complementary class, is identified when neural networks classify input images. The presence of human-recognizable information in adversarial perturbations allows researchers to gain insight into the working principles of neural networks and may lead to the development of techniques for detecting and defending against adversarial attacks.
♻ ☆ Video frame interpolation for high dynamic range sequences captured with dual-exposure sensors
Video frame interpolation (VFI) enables many important applications that might involve the temporal domain, such as slow motion playback, or the spatial domain, such as stop motion sequences. We are focusing on the former task, where one of the key challenges is handling high dynamic range (HDR) scenes in the presence of complex motion. To this end, we explore possible advantages of dual-exposure sensors that readily provide sharp short and blurry long exposures that are spatially registered and whose ends are temporally aligned. This way, motion blur registers temporally continuous information on the scene motion that, combined with the sharp reference, enables more precise motion sampling within a single camera shot. We demonstrate that this facilitates a more complex motion reconstruction in the VFI task, as well as HDR frame reconstruction that so far has been considered only for the originally captured frames, not in-between interpolated frames. We design a neural network trained in these tasks that clearly outperforms existing solutions. We also propose a metric for scene motion complexity that provides important insights into the performance of VFI methods at the test time.
comment: 17 pages, 10 figures
♻ ☆ GOOD: Exploring Geometric Cues for Detecting Objects in an Open World ICLR 2023
We address the task of open-world class-agnostic object detection, i.e., detecting every object in an image by learning from a limited number of base object classes. State-of-the-art RGB-based models suffer from overfitting the training classes and often fail at detecting novel-looking objects. This is because RGB-based models primarily rely on appearance similarity to detect novel objects and are also prone to overfitting short-cut cues such as textures and discriminative parts. To address these shortcomings of RGB-based object detectors, we propose incorporating geometric cues such as depth and normals, predicted by general-purpose monocular estimators. Specifically, we use the geometric cues to train an object proposal network for pseudo-labeling unannotated novel objects in the training set. Our resulting Geometry-guided Open-world Object Detector (GOOD) significantly improves detection recall for novel object categories and already performs well with only a few training classes. Using a single "person" class for training on the COCO dataset, GOOD surpasses SOTA methods by 5.0% AR@100, a relative improvement of 24%.
comment: Published as a conference paper at ICLR 2023
♻ ☆ Spotlights: Probing Shapes from Spherical Viewpoints ACCV2022
Recent years have witnessed the surge of learned representations that directly build upon point clouds. Though becoming increasingly expressive, most existing representations still struggle to generate ordered point sets. Inspired by spherical multi-view scanners, we propose a novel sampling model called Spotlights to represent a 3D shape as a compact 1D array of depth values. It simulates the configuration of cameras evenly distributed on a sphere, where each virtual camera casts light rays from its principal point through sample points on a small concentric spherical cap to probe for the possible intersections with the object surrounded by the sphere. The structured point cloud is hence given implicitly as a function of depths. We provide a detailed geometric analysis of this new sampling scheme and prove its effectiveness in the context of the point cloud completion task. Experimental results on both synthetic and real data demonstrate that our method achieves competitive accuracy and consistency while having a significantly reduced computational cost. Furthermore, we show superior performance on the downstream point cloud registration task over state-of-the-art completion methods.
comment: accepted by ACCV2022
♻ ☆ M3FAS: An Accurate and Robust MultiModal Mobile Face Anti-Spoofing System
Face presentation attacks (FPA), also known as face spoofing, have brought increasing concerns to the public through various malicious applications, such as financial fraud and privacy leakage. Therefore, safeguarding face recognition systems against FPA is of utmost importance. Although existing learning-based face anti-spoofing (FAS) models can achieve outstanding detection performance, they lack generalization capability and suffer significant performance drops in unforeseen environments. Many methodologies seek to use auxiliary modality data (e.g., depth and infrared maps) during the presentation attack detection (PAD) to address this limitation. However, these methods can be limited since (1) they require specific sensors such as depth and infrared cameras for data capture, which are rarely available on commodity mobile devices, and (2) they cannot work properly in practical scenarios when either modality is missing or of poor quality. In this paper, we devise an accurate and robust MultiModal Mobile Face Anti-Spoofing system named M3FAS to overcome the issues above. The innovation of this work mainly lies in the following aspects: (1) To achieve robust PAD, our system combines visual and auditory modalities using three pervasively available sensors: camera, speaker, and microphone; (2) We design a novel two-branch neural network with three hierarchical feature aggregation modules to perform cross-modal feature fusion; (3). We propose a multi-head training strategy. The model outputs three predictions from the vision, acoustic, and fusion heads, enabling a more flexible PAD. Extensive experiments have demonstrated the accuracy, robustness, and flexibility of M3FAS under various challenging experimental settings.
♻ ☆ Face Swapping as A Simple Arithmetic Operation
We propose a novel high-fidelity face swapping method called "Arithmetic Face Swapping" (AFS) that explicitly disentangles the intermediate latent space W+ of a pretrained StyleGAN into the "identity" and "style" subspaces so that a latent code in W+ is the sum of an "identity" code and a "style" code in the corresponding subspaces. Via our disentanglement, face swapping (FS) can be regarded as a simple arithmetic operation in W+, i.e., the summation of a source "identity" code and a target "style" code. This makes AFS more intuitive and elegant than other FS methods. In addition, our method can generalize over the standard face swapping to support other interesting operations, e.g., combining the identity of one source with styles of multiple targets and vice versa. We implement our identity-style disentanglement by learning a neural network that maps a latent code to a "style" code. We provide a condition for this network which theoretically guarantees identity preservation of the source face even after a sequence of face swapping operations. Extensive experiments demonstrate the advantage of our method over state-of-the-art FS methods in producing high-quality swapped faces. Our source code was made public at https://github.com/truongvu2000nd/AFS
♻ ☆ Language Quantized AutoEncoders: Towards Unsupervised Text-Image Alignment
Recent progress in scaling up large language models has shown impressive capabilities in performing few-shot learning across a wide range of text-based tasks. However, a key limitation is that these language models fundamentally lack visual perception - a crucial attribute needed to extend these models to be able to interact with the real world and solve vision tasks, such as in visual-question answering and robotics. Prior works have largely connected image to text through pretraining and/or fine-tuning on curated image-text datasets, which can be a costly and expensive process. In order to resolve this limitation, we propose a simple yet effective approach called Language-Quantized AutoEncoder (LQAE), a modification of VQ-VAE that learns to align text-image data in an unsupervised manner by leveraging pretrained language models (e.g., BERT, RoBERTa). Our main idea is to encode image as sequences of text tokens by directly quantizing image embeddings using a pretrained language codebook. We then apply random masking followed by a BERT model, and have the decoder reconstruct the original image from BERT predicted text token embeddings. By doing so, LQAE learns to represent similar images with similar clusters of text tokens, thereby aligning these two modalities without the use of aligned text-image pairs. This enables few-shot image classification with large language models (e.g., GPT-3) as well as linear classification of images based on BERT text features. To the best of our knowledge, our work is the first work that uses unaligned images for multimodal tasks by leveraging the power of pretrained language models.
comment: Fixed typos
♻ ☆ CPSeg: Cluster-free Panoptic Segmentation of 3D LiDAR Point Clouds ICRA 2023
A fast and accurate panoptic segmentation system for LiDAR point clouds is crucial for autonomous driving vehicles to understand the surrounding objects and scenes. Existing approaches usually rely on proposals or clustering to segment foreground instances. As a result, they struggle to achieve real-time performance. In this paper, we propose a novel real-time end-to-end panoptic segmentation network for LiDAR point clouds, called CPSeg. In particular, CPSeg comprises a shared encoder, a dual-decoder, and a cluster-free instance segmentation head, which is able to dynamically pillarize foreground points according to the learned embedding. Then, it acquires instance labels by finding connected pillars with a pairwise embedding comparison. Thus, the conventional proposal-based or clustering-based instance segmentation is transformed into a binary segmentation problem on the pairwise embedding comparison matrix. To help the network regress instance embedding, a fast and deterministic depth completion algorithm is proposed to calculate the surface normal of each point cloud in real-time. The proposed method is benchmarked on two large-scale autonomous driving datasets: SemanticKITTI and nuScenes. Notably, extensive experimental results show that CPSeg achieves state-of-the-art results among real-time approaches on both datasets.
comment: Accepted at ICRA 2023
Information Retrieval 13
☆ A Case Study for Compliance as Code with Graphs and Language Models: Public release of the Regulatory Knowledge Graph
The paper presents a study on using language models to automate the construction of executable Knowledge Graph (KG) for compliance. The paper focuses on Abu Dhabi Global Market regulations and taxonomy, involves manual tagging a portion of the regulations, training BERT-based models, which are then applied to the rest of the corpus. Coreference resolution and syntax analysis were used to parse the relationships between the tagged entities and to form KG stored in a Neo4j database. The paper states that the use of machine learning models released by regulators to automate the interpretation of rules is a vital step towards compliance automation, demonstrates the concept querying with Cypher, and states that the produced sub-graphs combined with Graph Neural Networks (GNN) will achieve expandability in judgment automation systems. The graph is open sourced on GitHub to provide structured data for future advancements in the field.
comment: 7 pages on RegKG step 1 in collaboration with ADGM
☆ Committed Private Information Retrieval
A private information retrieval (PIR) scheme allows a client to retrieve a data item $x_i$ among $n$ items $x_1,x_2,...,x_n$ from $k$ servers, without revealing what $i$ is even when $t < k$ servers collude and try to learn $i$. Such a PIR scheme is said to be $t$-private. A PIR scheme is $v$-verifiable if the client can verify the correctness of the retrieved $x_i$ even when $v \leq k$ servers collude and try to fool the client by sending manipulated data. Most of the previous works in the literature on PIR assumed that $v < k$, leaving the case of all-colluding servers open. We propose a generic construction that combines a linear map commitment (LMC) and an arbitrary linear PIR scheme to produce a $k$-verifiable PIR scheme, termed a committed PIR scheme. Such a scheme guarantees that even in the worst scenario, when all servers are under the control of an attacker, although the privacy is unavoidably lost, the client won't be fooled into accepting an incorrect $x_i$. We demonstrate the practicality of our proposal by implementing the committed PIR schemes based on the Lai-Malavolta LMC and three well-known PIR schemes using the GMP library and \texttt{blst}, the current fastest C library for elliptic curve pairings.
☆ Reinforcing User Retention in a Billion Scale Short Video Recommender System
Recently, short video platforms have achieved rapid user growth by recommending interesting content to users. The objective of the recommendation is to optimize user retention, thereby driving the growth of DAU (Daily Active Users). Retention is a long-term feedback after multiple interactions of users and the system, and it is hard to decompose retention reward to each item or a list of items. Thus traditional point-wise and list-wise models are not able to optimize retention. In this paper, we choose reinforcement learning methods to optimize the retention as they are designed to maximize the long-term performance. We formulate the problem as an infinite-horizon request-based Markov Decision Process, and our objective is to minimize the accumulated time interval of multiple sessions, which is equal to improving the app open frequency and user retention. However, current reinforcement learning algorithms can not be directly applied in this setting due to uncertainty, bias, and long delay time incurred by the properties of user retention. We propose a novel method, dubbed RLUR, to address the aforementioned challenges. Both offline and live experiments show that RLUR can significantly improve user retention. RLUR has been fully launched in Kuaishou app for a long time, and achieves consistent performance improvement on user retention and DAU.
☆ Two-Stage Constrained Actor-Critic fo Short Video Recommendation
The wide popularity of short videos on social media poses new opportunities and challenges to optimize recommender systems on the video-sharing platforms. Users sequentially interact with the system and provide complex and multi-faceted responses, including watch time and various types of interactions with multiple videos. One the one hand, the platforms aims at optimizing the users' cumulative watch time (main goal) in long term, which can be effectively optimized by Reinforcement Learning. On the other hand, the platforms also needs to satisfy the constraint of accommodating the responses of multiple user interactions (auxiliary goals) such like, follow, share etc. In this paper, we formulate the problem of short video recommendation as a Constrained Markov Decision Process (CMDP). We find that traditional constrained reinforcement learning algorithms can not work well in this setting. We propose a novel two-stage constrained actor-critic method: At stage one, we learn individual policies to optimize each auxiliary signal. At stage two, we learn a policy to (i) optimize the main signal and (ii) stay close to policies learned at the first stage, which effectively guarantees the performance of this main policy on the auxiliaries. Through extensive offline evaluations, we demonstrate effectiveness of our method over alternatives in both optimizing the main goal as well as balancing the others. We further show the advantage of our method in live experiments of short video recommendations, where it significantly outperforms other baselines in terms of both watch time and interactions. Our approach has been fully launched in the production system to optimize user experiences on the platform.
comment: arXiv admin note: substantial text overlap with arXiv:2205.13248
☆ Modeling Sequential Sentence Relation to Improve Cross-lingual Dense Retrieval ICLR 2023
Recently multi-lingual pre-trained language models (PLM) such as mBERT and XLM-R have achieved impressive strides in cross-lingual dense retrieval. Despite its successes, they are general-purpose PLM while the multilingual PLM tailored for cross-lingual retrieval is still unexplored. Motivated by an observation that the sentences in parallel documents are approximately in the same order, which is universal across languages, we propose to model this sequential sentence relation to facilitate cross-lingual representation learning. Specifically, we propose a multilingual PLM called masked sentence model (MSM), which consists of a sentence encoder to generate the sentence representations, and a document encoder applied to a sequence of sentence vectors from a document. The document encoder is shared for all languages to model the universal sequential sentence relation across languages. To train the model, we propose a masked sentence prediction task, which masks and predicts the sentence vector via a hierarchical contrastive loss with sampled negatives. Comprehensive experiments on four cross-lingual retrieval tasks show MSM significantly outperforms existing advanced pre-training models, demonstrating the effectiveness and stronger cross-lingual retrieval capabilities of our approach. Code and model will be available.
comment: Published at ICLR 2023
☆ Improving Recommendation Relevance by simulating User Interest
Most if not all on-line item-to-item recommendation systems rely on estimation of a distance like measure (rank) of similarity between items. For on-line recommendation systems, time sensitivity of this similarity measure is extremely important. We observe that recommendation "recency" can be straightforwardly and transparently maintained by iterative reduction of ranks of inactive items. The paper briefly summarizes algorithmic developments based on this self-explanatory observation. The basic idea behind this work is patented in a context of online recommendation systems.
☆ ANTM: An Aligned Neural Topic Model for Exploring Evolving Topics
As the amount of text data generated by humans and machines increases, the necessity of understanding large corpora and finding a way to extract insights from them is becoming more crucial than ever. Dynamic topic models are effective methods that primarily focus on studying the evolution of topics present in a collection of documents. These models are widely used for understanding trends, exploring public opinion in social networks, or tracking research progress and discoveries in scientific archives. Since topics are defined as clusters of semantically similar documents, it is necessary to observe the changes in the content or themes of these clusters in order to understand how topics evolve as new knowledge is discovered over time. In this paper, we introduce the Aligned Neural Topic Model (ANTM), a dynamic neural topic model that uses document embeddings to compute clusters of semantically similar documents at different periods and to align document clusters to represent their evolution. This alignment procedure preserves the temporal similarity of document clusters over time and captures the semantic change of words characterized by their context within different periods. Experiments on four different datasets show that ANTM outperforms probabilistic dynamic topic models (e.g. DTM, DETM) and significantly improves topic coherence and diversity over other existing dynamic neural topic models (e.g. BERTopic).
☆ How Bad is Top-$K$ Recommendation under Competing Content Creators?
Content creators compete for exposure on recommendation platforms, and such strategic behavior leads to a dynamic shift over the content distribution. However, how the creators' competition impacts user welfare and how the relevance-driven recommendation influences the dynamics in the long run are still largely unknown. This work provides theoretical insights into these research questions. We model the creators' competition under the assumptions that: 1) the platform employs an innocuous top-$K$ recommendation policy; 2) user decisions follow the Random Utility model; 3) content creators compete for user engagement and, without knowing their utility function in hindsight, apply arbitrary no-regret learning algorithms to update their strategies. We study the user welfare guarantee through the lens of Price of Anarchy and show that the fraction of user welfare loss due to creator competition is always upper bounded by a small constant depending on $K$ and randomness in user decisions; we also prove the tightness of this bound. Our result discloses an intrinsic merit of the myopic approach to the recommendation, i.e., relevance-driven matching performs reasonably well in the long run, as long as users' decisions involve randomness and the platform provides reasonably many alternatives to its users.
☆ Can ChatGPT Write a Good Boolean Query for Systematic Review Literature Search?
Systematic reviews are comprehensive reviews of the literature for a highly focused research question. These reviews are often treated as the highest form of evidence in evidence-based medicine, and are the key strategy to answer research questions in the medical field. To create a high-quality systematic review, complex Boolean queries are often constructed to retrieve studies for the review topic. However, it often takes a long time for systematic review researchers to construct a high quality systematic review Boolean query, and often the resulting queries are far from effective. Poor queries may lead to biased or invalid reviews, because they missed to retrieve key evidence, or to extensive increase in review costs, because they retrieved too many irrelevant studies. Recent advances in Transformer-based generative models have shown great potential to effectively follow instructions from users and generate answers based on the instructions being made. In this paper, we investigate the effectiveness of the latest of such models, ChatGPT, in generating effective Boolean queries for systematic review literature search. Through a number of extensive experiments on standard test collections for the task, we find that ChatGPT is capable of generating queries that lead to high search precision, although trading-off this for recall. Overall, our study demonstrates the potential of ChatGPT in generating effective Boolean queries for systematic review literature search. The ability of ChatGPT to follow complex instructions and generate queries with high precision makes it a valuable tool for researchers conducting systematic reviews, particularly for rapid reviews where time is a constraint and often trading-off higher precision for lower recall is acceptable.
♻ ☆ Locality Sensitive Hashing for Structured Data: A Survey
Data similarity (or distance) computation is a fundamental research topic which fosters a variety of similarity-based machine learning and data mining applications. In big data analytics, it is impractical to compute the exact similarity of data instances due to high computational cost. To this end, the Locality Sensitive Hashing (LSH) technique has been proposed to provide accurate estimators for various similarity measures between sets or vectors in an efficient manner without the learning process. Structured data (e.g., sequences, trees and graphs), which are composed of elements and relations between the elements, are commonly seen in the real world, but the traditional LSH algorithms cannot preserve the structure information represented as relations between elements. In order to conquer the issue, researchers have been devoted to the family of the hierarchical LSH algorithms. In this paper, we explore the present progress of the research into hierarchical LSH from the following perspectives: 1) Data structures, where we review various hierarchical LSH algorithms for three typical data structures and uncover their inherent connections; 2) Applications, where we review the hierarchical LSH algorithms in multiple application scenarios; 3) Challenges, where we discuss some potential challenges as future directions.
♻ ☆ Towards Disentangling Relevance and Bias in Unbiased Learning to Rank
Unbiased learning to rank (ULTR) studies the problem of mitigating various biases from implicit user feedback data such as clicks, and has been receiving considerable attention recently. A popular ULTR approach for real-world applications uses a two-tower architecture, where click modeling is factorized into a relevance tower with regular input features, and a bias tower with bias-relevant inputs such as the position of a document. A successful factorization will allow the relevance tower to be exempt from biases. In this work, we identify a critical issue that existing ULTR methods ignored - the bias tower can be confounded with the relevance tower via the underlying true relevance. In particular, the positions were determined by the logging policy, i.e., the previous production model, which would possess relevance information. We give both theoretical analysis and empirical results to show the negative effects on relevance tower due to such a correlation. We then propose three methods to mitigate the negative confounding effects by better disentangling relevance and bias. Empirical results on both controlled public datasets and a large-scale industry dataset show the effectiveness of the proposed approaches.
♻ ☆ Evaluation Metrics for Measuring Bias in Search Engine Results
Search engines decide what we see for a given search query. Since many people are exposed to information through search engines, it is fair to expect that search engines are neutral. However, search engine results do not necessarily cover all the viewpoints of a search query topic, and they can be biased towards a specific view since search engine results are returned based on relevance, which is calculated using many features and sophisticated algorithms where search neutrality is not necessarily the focal point. Therefore, it is important to evaluate the search engine results with respect to bias. In this work we propose novel web search bias evaluation measures which take into account the rank and relevance. We also propose a framework to evaluate web search bias using the proposed measures and test our framework on two popular search engines based on 57 controversial query topics such as abortion, medical marijuana, and gay marriage. We measure the stance bias (in support or against), as well as the ideological bias (conservative or liberal). We observe that the stance does not necessarily correlate with the ideological leaning, e.g. a positive stance on abortion indicates a liberal leaning but a positive stance on Cuba embargo indicates a conservative leaning. Our experiments show that neither of the search engines suffers from stance bias. However, both search engines suffer from ideological bias, both favouring one ideological leaning to the other, which is more significant from the perspective of polarisation in our society.
♻ ☆ Experimenting with Selected Automated Approaches for Bias Analysis
This work first presents our attempts to establish an automated model using state-of-the-art approaches for analysing bias in search results of Bing and Google. Experimental results indicate that the current class-wise F1-scores of our best model are not sufficient to establish an automated model for bias analysis. Thus, we decided not to continue with this approach.
Machine Learning 132
☆ Aligning Robot and Human Representations
To act in the world, robots rely on a representation of salient task aspects: for example, to carry a cup of coffee, a robot must consider movement efficiency and cup orientation in its behaviour. However, if we want robots to act for and with people, their representations must not be just functional but also reflective of what humans care about, i.e. their representations must be aligned with humans'. In this survey, we pose that current reward and imitation learning approaches suffer from representation misalignment, where the robot's learned representation does not capture the human's representation. We suggest that because humans will be the ultimate evaluator of robot performance in the world, it is critical that we explicitly focus our efforts on aligning learned task representations with humans, in addition to learning the downstream task. We advocate that current representation learning approaches in robotics should be studied from the perspective of how well they accomplish the objective of representation alignment. To do so, we mathematically define the problem, identify its key desiderata, and situate current robot learning methods within this formalism. We conclude the survey by suggesting future directions for exploring open challenges.
comment: 19 pages, 1 figure, 1 table
☆ Learning a Fourier Transform for Linear Relative Positional Encodings in Transformers
We propose a new class of linear Transformers called FourierLearner-Transformers (FLTs), which incorporate a wide range of relative positional encoding mechanisms (RPEs). These include regular RPE techniques applied for nongeometric data, as well as novel RPEs operating on the sequences of tokens embedded in higher-dimensional Euclidean spaces (e.g. point clouds). FLTs construct the optimal RPE mechanism implicitly by learning its spectral representation. As opposed to other architectures combining efficient low-rank linear attention with RPEs, FLTs remain practical in terms of their memory usage and do not require additional assumptions about the structure of the RPE-mask. FLTs allow also for applying certain structural inductive bias techniques to specify masking strategies, e.g. they provide a way to learn the so-called local RPEs introduced in this paper and providing accuracy gains as compared with several other linear Transformers for language modeling. We also thoroughly tested FLTs on other data modalities and tasks, such as: image classification and 3D molecular modeling. For 3D-data FLTs are, to the best of our knowledge, the first Transformers architectures providing RPE-enhanced linear attention.
☆ PyGlove: Efficiently Exchanging ML Ideas as Code
The increasing complexity and scale of machine learning (ML) has led to the need for more efficient collaboration among multiple teams. For example, when a research team invents a new architecture like "ResNet," it is desirable for multiple engineering teams to adopt it. However, the effort required for each team to study and understand the invention does not scale well with the number of teams or inventions. In this paper, we present an extension of our PyGlove library to easily and scalably share ML ideas. PyGlove represents ideas as symbolic rule-based patches, enabling researchers to write down the rules for models they have not seen. For example, an inventor can write rules that will "add skip-connections." This permits a network effect among teams: at once, any team can issue patches to all other teams. Such a network effect allows users to quickly surmount the cost of adopting PyGlove by writing less code quicker, providing a benefit that scales with time. We describe the new paradigm of organizing ML through symbolic patches and compare it to existing approaches. We also perform a case study of a large codebase where PyGlove led to an 80% reduction in the number of lines of code.
comment: 8 pages, 10 figures, 1 table
☆ Enhancing Once-For-All: A Study on Parallel Blocks, Skip Connections and Early Exits
The use of Neural Architecture Search (NAS) techniques to automate the design of neural networks has become increasingly popular in recent years. The proliferation of devices with different hardware characteristics using such neural networks, as well as the need to reduce the power consumption for their search, has led to the realisation of Once-For-All (OFA), an eco-friendly algorithm characterised by the ability to generate easily adaptable models through a single learning process. In order to improve this paradigm and develop high-performance yet eco-friendly NAS techniques, this paper presents OFAv2, the extension of OFA aimed at improving its performance while maintaining the same ecological advantage. The algorithm is improved from an architectural point of view by including early exits, parallel blocks and dense skip connections. The training process is extended by two new phases called Elastic Level and Elastic Height. A new Knowledge Distillation technique is presented to handle multi-output networks, and finally a new strategy for dynamic teacher network selection is proposed. These modifications allow OFAv2 to improve its accuracy performance on the Tiny ImageNet dataset by up to 12.07% compared to the original version of OFA, while maintaining the algorithm flexibility and advantages.
☆ Analyzing the impact of climate change on critical infrastructure from the scientific literature: A weakly supervised NLP approach
Natural language processing (NLP) is a promising approach for analyzing large volumes of climate-change and infrastructure-related scientific literature. However, best-in-practice NLP techniques require large collections of relevant documents (corpus). Furthermore, NLP techniques using machine learning and deep learning techniques require labels grouping the articles based on user-defined criteria for a significant subset of a corpus in order to train the supervised model. Even labeling a few hundred documents with human subject-matter experts is a time-consuming process. To expedite this process, we developed a weak supervision-based NLP approach that leverages semantic similarity between categories and documents to (i) establish a topic-specific corpus by subsetting a large-scale open-access corpus and (ii) generate category labels for the topic-specific corpus. In comparison with a months-long process of subject-matter expert labeling, we assign category labels to the whole corpus using weak supervision and supervised learning in about 13 hours. The labeled climate and NCF corpus enable targeted, efficient identification of documents discussing a topic (or combination of topics) of interest and identification of various effects of climate change on critical infrastructure, improving the usability of scientific literature and ultimately supporting enhanced policy and decision making. To demonstrate this capability, we conduct topic modeling on pairs of climate hazards and NCFs to discover trending topics at the intersection of these categories. This method is useful for analysts and decision-makers to quickly grasp the relevant topics and most important documents linked to the topic.
☆ AdaptDiffuser: Diffusion Models as Adaptive Self-evolving Planners
Diffusion models have demonstrated their powerful generative capability in many tasks, with great potential to serve as a paradigm for offline reinforcement learning. However, the quality of the diffusion model is limited by the insufficient diversity of training data, which hinders the performance of planning and the generalizability to new tasks. This paper introduces AdaptDiffuser, an evolutionary planning method with diffusion that can self-evolve to improve the diffusion model hence a better planner, not only for seen tasks but can also adapt to unseen tasks. AdaptDiffuser enables the generation of rich synthetic expert data for goal-conditioned tasks using guidance from reward gradients. It then selects high-quality data via a discriminator to finetune the diffusion model, which improves the generalization ability to unseen tasks. Empirical experiments on two benchmark environments and two carefully designed unseen tasks in KUKA industrial robot arm and Maze2D environments demonstrate the effectiveness of AdaptDiffuser. For example, AdaptDiffuser not only outperforms the previous art Diffuser by 20.8% on Maze2D and 7.5% on MuJoCo locomotion, but also adapts better to new tasks, e.g., KUKA pick-and-place, by 27.9% without requiring additional expert data.
☆ From Robustness to Privacy and Back
We study the relationship between two desiderata of algorithms in statistical inference and machine learning: differential privacy and robustness to adversarial data corruptions. Their conceptual similarity was first observed by Dwork and Lei (STOC 2009), who observed that private algorithms satisfy robustness, and gave a general method for converting robust algorithms to private ones. However, all general methods for transforming robust algorithms into private ones lead to suboptimal error rates. Our work gives the first black-box transformation that converts any adversarially robust algorithm into one that satisfies pure differential privacy. Moreover, we show that for any low-dimensional estimation task, applying our transformation to an optimal robust estimator results in an optimal private estimator. Thus, we conclude that for any low-dimensional task, the optimal error rate for $\varepsilon$-differentially private estimators is essentially the same as the optimal error rate for estimators that are robust to adversarially corrupting $1/\varepsilon$ training samples. We apply our transformation to obtain new optimal private estimators for several high-dimensional tasks, including Gaussian (sparse) linear regression and PCA. Finally, we present an extension of our transformation that leads to approximate differentially private algorithms whose error does not depend on the range of the output space, which is impossible under pure differential privacy.
☆ Unsupervised hierarchical clustering using the learning dynamics of RBMs
Datasets in the real world are often complex and to some degree hierarchical, with groups and sub-groups of data sharing common characteristics at different levels of abstraction. Understanding and uncovering the hidden structure of these datasets is an important task that has many practical applications. To address this challenge, we present a new and general method for building relational data trees by exploiting the learning dynamics of the Restricted Boltzmann Machine (RBM). Our method is based on the mean-field approach, derived from the Plefka expansion, and developed in the context of disordered systems. It is designed to be easily interpretable. We tested our method in an artificially created hierarchical dataset and on three different real-world datasets (images of digits, mutations in the human genome, and a homologous family of proteins). The method is able to automatically identify the hierarchical structure of the data. This could be useful in the study of homologous protein sequences, where the relationships between proteins are critical for understanding their function and evolution.
comment: 33 pages, 17 figures
☆ A Case Study for Compliance as Code with Graphs and Language Models: Public release of the Regulatory Knowledge Graph
The paper presents a study on using language models to automate the construction of executable Knowledge Graph (KG) for compliance. The paper focuses on Abu Dhabi Global Market regulations and taxonomy, involves manual tagging a portion of the regulations, training BERT-based models, which are then applied to the rest of the corpus. Coreference resolution and syntax analysis were used to parse the relationships between the tagged entities and to form KG stored in a Neo4j database. The paper states that the use of machine learning models released by regulators to automate the interpretation of rules is a vital step towards compliance automation, demonstrates the concept querying with Cypher, and states that the produced sub-graphs combined with Graph Neural Networks (GNN) will achieve expandability in judgment automation systems. The graph is open sourced on GitHub to provide structured data for future advancements in the field.
comment: 7 pages on RegKG step 1 in collaboration with ADGM
☆ Coinductive guide to inductive transformer heads
We argue that all building blocks of transformer models can be expressed with a single concept: combinatorial Hopf algebra. Transformer learning emerges as a result of the subtle interplay between the algebraic and coalgebraic operations of the combinatorial Hopf algebra. Viewed through this lens, the transformer model becomes a linear time-invariant system where the attention mechanism computes a generalized convolution transform and the residual stream serves as a unit impulse. Attention-only transformers then learn by enforcing an invariant between these two paths. We call this invariant Hopf coherence. Due to this, with a degree of poetic license, one could call combinatorial Hopf algebras "tensors with a built-in loss function gradient". This loss function gradient occurs within the single layers and no backward pass is needed. This is in contrast to automatic differentiation which happens across the whole graph and needs a explicit backward pass. This property is the result of the fact that combinatorial Hopf algebras have the surprising property of calculating eigenvalues by repeated squaring.
☆ Online Ad Allocation with Predictions
Display Ads and the generalized assignment problem are two well-studied online packing problems with important applications in ad allocation and other areas. In both problems, ad impressions arrive online and have to be allocated immediately to budget-constrained advertisers. Worst-case algorithms that achieve the ideal competitive ratio are known, but might act overly conservative given the predictable and usually tame nature of real-world input. Given this discrepancy, we develop an algorithm for both problems that incorporate machine-learned predictions and can thus improve the performance beyond the worst-case. Our algorithm is based on the work of Feldman et al. (2009) and similar in nature to Mahdian et al. (2007) who were the first to develop a learning-augmented algorithm for the related, but more structured Ad Words problem. We use a novel analysis to show that our algorithm is able to capitalize on a good prediction, while being robust against poor predictions. We experimentally evaluate our algorithm on synthetic and real-world data on a wide range of predictions. Our algorithm is consistently outperforming the worst-case algorithm without predictions.
☆ Leveraging weak complementary labels to improve semantic segmentation of hepatocellular carcinoma and cholangiocarcinoma in H&E-stained slides
In this paper, we present a deep learning segmentation approach to classify and quantify the two most prevalent primary liver cancers - hepatocellular carcinoma and intrahepatic cholangiocarcinoma - from hematoxylin and eosin (H&E) stained whole slide images. While semantic segmentation of medical images typically requires costly pixel-level annotations by domain experts, there often exists additional information which is routinely obtained in clinical diagnostics but rarely utilized for model training. We propose to leverage such weak information from patient diagnoses by deriving complementary labels that indicate to which class a sample cannot belong to. To integrate these labels, we formulate a complementary loss for segmentation. Motivated by the medical application, we demonstrate for general segmentation tasks that including additional patches with solely weak complementary labels during model training can significantly improve the predictive performance and robustness of a model. On the task of diagnostic differentiation between hepatocellular carcinoma and intrahepatic cholangiocarcinoma, we achieve a balanced accuracy of 0.91 (CI 95%: 0.86 - 0.95) at case level for 165 hold-out patients. Furthermore, we also show that leveraging complementary labels improves the robustness of segmentation and increases performance at case level.
☆ PINN Training using Biobjective Optimization: The Trade-off between Data Loss and Residual Loss
Physics informed neural networks (PINNs) have proven to be an efficient tool to represent problems for which measured data are available and for which the dynamics in the data are expected to follow some physical laws. In this paper, we suggest a multiobjective perspective on the training of PINNs by treating the data loss and the residual loss as two individual objective functions in a truly biobjective optimization approach. As a showcase example, we consider COVID-19 predictions in Germany and built an extended susceptibles-infected-recovered (SIR) model with additionally considered leaky-vaccinated and hospitalized populations (SVIHR model) to model the transition rates and to predict future infections. SIR-type models are expressed by systems of ordinary differential equations (ODEs). We investigate the suitability of the generated PINN for COVID-19 predictions and compare the resulting predicted curves with those obtained by applying the method of non-standard finite differences to the system of ODEs and initial data. The approach is applicable to various systems of ODEs that define dynamical regimes. Those regimes do not need to be SIR-type models, and the corresponding underlying data sets do not have to be associated with COVID-19.
comment: 47 pages
☆ Creating Probabilistic Forecasts from Arbitrary Deterministic Forecasts using Conditional Invertible Neural Networks
In various applications, probabilistic forecasts are required to quantify the inherent uncertainty associated with the forecast. However, numerous modern forecasting methods are still designed to create deterministic forecasts. Transforming these deterministic forecasts into probabilistic forecasts is often challenging and based on numerous assumptions that may not hold in real-world situations. Therefore, the present article proposes a novel approach for creating probabilistic forecasts from arbitrary deterministic forecasts. In order to implement this approach, we use a conditional Invertible Neural Network (cINN). More specifically, we apply a cINN to learn the underlying distribution of the data and then combine the uncertainty from this distribution with an arbitrary deterministic forecast to generate accurate probabilistic forecasts. Our approach enables the simple creation of probabilistic forecasts without complicated statistical loss functions or further assumptions. Besides showing the mathematical validity of our approach, we empirically show that our approach noticeably outperforms traditional methods for including uncertainty in deterministic forecasts and generally outperforms state-of-the-art probabilistic forecasting benchmarks.
comment: Preprint submitted to the International Journal of Forecasting
☆ Interpretations of Domain Adaptations via Layer Variational Analysis ICLR 2023
Transfer learning is known to perform efficiently in many applications empirically, yet limited literature reports the mechanism behind the scene. This study establishes both formal derivations and heuristic analysis to formulate the theory of transfer learning in deep learning. Our framework utilizing layer variational analysis proves that the success of transfer learning can be guaranteed with corresponding data conditions. Moreover, our theoretical calculation yields intuitive interpretations towards the knowledge transfer process. Subsequently, an alternative method for network-based transfer learning is derived. The method shows an increase in efficiency and accuracy for domain adaptation. It is particularly advantageous when new domain data is sufficiently sparse during adaptation. Numerical experiments over diverse tasks validated our theory and verified that our analytic expression achieved better performance in domain adaptation than the gradient descent method.
comment: To be published at ICLR 2023
☆ Fixing by Mixing: A Recipe for Optimal Byzantine ML under Heterogeneity AISTATS 2023
Byzantine machine learning (ML) aims to ensure the resilience of distributed learning algorithms to misbehaving (or Byzantine) machines. Although this problem received significant attention, prior works often assume the data held by the machines to be homogeneous, which is seldom true in practical settings. Data heterogeneity makes Byzantine ML considerably more challenging, since a Byzantine machine can hardly be distinguished from a non-Byzantine outlier. A few solutions have been proposed to tackle this issue, but these provide suboptimal probabilistic guarantees and fare poorly in practice. This paper closes the theoretical gap, achieving optimality and inducing good empirical results. In fact, we show how to automatically adapt existing solutions for (homogeneous) Byzantine ML to the heterogeneous setting through a powerful mechanism, we call nearest neighbor mixing (NNM), which boosts any standard robust distributed gradient descent variant to yield optimal Byzantine resilience under heterogeneity. We obtain similar guarantees (in expectation) by plugging NNM in the distributed stochastic heavy ball method, a practical substitute to distributed gradient descent. We obtain empirical results that significantly outperform state-of-the-art Byzantine ML solutions.
comment: Accepted paper at AISTATS 2023
☆ Using Explainability to Inform Statistical Downscaling Based on Deep Learning Beyond Standard Validation Approaches
Deep learning (DL) has emerged as a promising tool to downscale climate projections at regional-to-local scales from large-scale atmospheric fields following the perfect-prognosis (PP) approach. Given their complexity, it is crucial to properly evaluate these methods, especially when applied to changing climatic conditions where the ability to extrapolate/generalise is key. In this work, we intercompare several DL models extracted from the literature for the same challenging use-case (downscaling temperature in the CORDEX North America domain) and expand standard evaluation methods building on eXplainable artifical intelligence (XAI) techniques. We show how these techniques can be used to unravel the internal behaviour of these models, providing new evaluation dimensions and aiding in their diagnostic and design. These results show the usefulness of incorporating XAI techniques into statistical downscaling evaluation frameworks, especially when working with large regions and/or under climate change conditions.
comment: Under review for Journal of Advances in Modeling Earth Systems
☆ A Systematic Evaluation of Backdoor Trigger Characteristics in Image Classification
Deep learning achieves outstanding results in many machine learning tasks. Nevertheless, it is vulnerable to backdoor attacks that modify the training set to embed a secret functionality in the trained model. The modified training samples have a secret property, i.e., a trigger. At inference time, the secret functionality is activated when the input contains the trigger, while the model functions correctly in other cases. While there are many known backdoor attacks (and defenses), deploying a stealthy attack is still far from trivial. Successfully creating backdoor triggers heavily depends on numerous parameters. Unfortunately, research has not yet determined which parameters contribute most to the attack performance. This paper systematically analyzes the most relevant parameters for the backdoor attacks, i.e., trigger size, position, color, and poisoning rate. Using transfer learning, which is very common in computer vision, we evaluate the attack on numerous state-of-the-art models (ResNet, VGG, AlexNet, and GoogLeNet) and datasets (MNIST, CIFAR10, and TinyImageNet). Our attacks cover the majority of backdoor settings in research, providing concrete directions for future works. Our code is publicly available to facilitate the reproducibility of our results.
☆ AIROGS: Artificial Intelligence for RObust Glaucoma Screening Challenge
The early detection of glaucoma is essential in preventing visual impairment. Artificial intelligence (AI) can be used to analyze color fundus photographs (CFPs) in a cost-effective manner, making glaucoma screening more accessible. While AI models for glaucoma screening from CFPs have shown promising results in laboratory settings, their performance decreases significantly in real-world scenarios due to the presence of out-of-distribution and low-quality images. To address this issue, we propose the Artificial Intelligence for Robust Glaucoma Screening (AIROGS) challenge. This challenge includes a large dataset of around 113,000 images from about 60,000 patients and 500 different screening centers, and encourages the development of algorithms that are robust to ungradable and unexpected input data. We evaluated solutions from 14 teams in this paper, and found that the best teams performed similarly to a set of 20 expert ophthalmologists and optometrists. The highest-scoring team achieved an area under the receiver operating characteristic curve of 0.99 (95% CI: 0.98-0.99) for detecting ungradable images on-the-fly. Additionally, many of the algorithms showed robust performance when tested on three other publicly available datasets. These results demonstrate the feasibility of robust AI-enabled glaucoma screening.
comment: 19 pages, 8 figures, 3 tables
☆ Rethinking Semi-Supervised Medical Image Segmentation: A Variance-Reduction Perspective
For medical image segmentation, contrastive learning is the dominant practice to improve the quality of visual representations by contrasting semantically similar and dissimilar pairs of samples. This is enabled by the observation that without accessing ground truth label, negative examples with truly dissimilar anatomical features, if sampled, can significantly improve the performance. In reality, however, these samples may come from similar anatomical features and the models may struggle to distinguish the minority tail-class samples, making the tail classes more prone to misclassification, both of which typically lead to model collapse. In this paper, we propose ARCO, a semi-supervised contrastive learning (CL) framework with stratified group sampling theory in medical image segmentation. In particular, we first propose building ARCO through the concept of variance-reduced estimation, and show that certain variance-reduction techniques are particularly beneficial in medical image segmentation tasks with extremely limited labels. Furthermore, we theoretically prove these sampling techniques are universal in variance reduction. Finally, we experimentally validate our approaches on three benchmark datasets with different label settings, and our methods consistently outperform state-of-the-art semi- and fully-supervised methods. Additionally, we augment the CL frameworks with these sampling techniques and demonstrate significant gains over previous methods. We believe our work is an important step towards semi-supervised medical image segmentation by quantifying the limitation of current self-supervision objectives for accomplishing medical image analysis tasks.
☆ Stochastic Policy Gradient Methods: Improved Sample Complexity for Fisher-non-degenerate Policies
Recently, the impressive empirical success of policy gradient (PG) methods has catalyzed the development of their theoretical foundations. Despite the huge efforts directed at the design of efficient stochastic PG-type algorithms, the understanding of their convergence to a globally optimal policy is still limited. In this work, we develop improved global convergence guarantees for a general class of Fisher-non-degenerate parameterized policies which allows to address the case of continuous state action spaces. First, we propose a Normalized Policy Gradient method with Implicit Gradient Transport (N-PG-IGT) and derive a $\tilde{\mathcal{O}}(\varepsilon^{-2.5})$ sample complexity of this method for finding a global $\varepsilon$-optimal policy. Improving over the previously known $\tilde{\mathcal{O}}(\varepsilon^{-3})$ complexity, this algorithm does not require the use of importance sampling or second-order information and samples only one trajectory per iteration. Second, we further improve this complexity to $\tilde{ \mathcal{\mathcal{O}} }(\varepsilon^{-2})$ by considering a Hessian-Aided Recursive Policy Gradient ((N)-HARPG) algorithm enhanced with a correction based on a Hessian-vector product. Interestingly, both algorithms are $(i)$ simple and easy to implement: single-loop, do not require large batches of trajectories and sample at most two trajectories per iteration; $(ii)$ computationally and memory efficient: they do not require expensive subroutines at each iteration and can be implemented with memory linear in the dimension of parameters.
comment: This work was initially submitted in October 2022
☆ Distributional constrained reinforcement learning for supply chain optimization
This work studies reinforcement learning (RL) in the context of multi-period supply chains subject to constraints, e.g., on production and inventory. We introduce Distributional Constrained Policy Optimization (DCPO), a novel approach for reliable constraint satisfaction in RL. Our approach is based on Constrained Policy Optimization (CPO), which is subject to approximation errors that in practice lead it to converge to infeasible policies. We address this issue by incorporating aspects of distributional RL into DCPO. Specifically, we represent the return and cost value functions using neural networks that output discrete distributions, and we reshape costs based on the associated confidence. Using a supply chain case study, we show that DCPO improves the rate at which the RL policy converges and ensures reliable constraint satisfaction by the end of training. The proposed method also improves predictability, greatly reducing the variance of returns between runs, respectively; this result is significant in the context of policy gradient methods, which intrinsically introduce significant variance during training.
comment: 6 pages, 4 figures
☆ Reinforcing User Retention in a Billion Scale Short Video Recommender System
Recently, short video platforms have achieved rapid user growth by recommending interesting content to users. The objective of the recommendation is to optimize user retention, thereby driving the growth of DAU (Daily Active Users). Retention is a long-term feedback after multiple interactions of users and the system, and it is hard to decompose retention reward to each item or a list of items. Thus traditional point-wise and list-wise models are not able to optimize retention. In this paper, we choose reinforcement learning methods to optimize the retention as they are designed to maximize the long-term performance. We formulate the problem as an infinite-horizon request-based Markov Decision Process, and our objective is to minimize the accumulated time interval of multiple sessions, which is equal to improving the app open frequency and user retention. However, current reinforcement learning algorithms can not be directly applied in this setting due to uncertainty, bias, and long delay time incurred by the properties of user retention. We propose a novel method, dubbed RLUR, to address the aforementioned challenges. Both offline and live experiments show that RLUR can significantly improve user retention. RLUR has been fully launched in Kuaishou app for a long time, and achieves consistent performance improvement on user retention and DAU.
☆ Leveraging Contaminated Datasets to Learn Clean-Data Distribution with Purified Generative Adversarial Networks
Generative adversarial networks (GANs) are known for their strong abilities on capturing the underlying distribution of training instances. Since the seminal work of GAN, many variants of GAN have been proposed. However, existing GANs are almost established on the assumption that the training dataset is clean. But in many real-world applications, this may not hold, that is, the training dataset may be contaminated by a proportion of undesired instances. When training on such datasets, existing GANs will learn a mixture distribution of desired and contaminated instances, rather than the desired distribution of desired data only (target distribution). To learn the target distribution from contaminated datasets, two purified generative adversarial networks (PuriGAN) are developed, in which the discriminators are augmented with the capability to distinguish between target and contaminated instances by leveraging an extra dataset solely composed of contamination instances. We prove that under some mild conditions, the proposed PuriGANs are guaranteed to converge to the distribution of desired instances. Experimental results on several datasets demonstrate that the proposed PuriGANs are able to generate much better images from the desired distribution than comparable baselines when trained on contaminated datasets. In addition, we also demonstrate the usefulness of PuriGAN on downstream applications by applying it to the tasks of semi-supervised anomaly detection on contaminated datasets and PU-learning. Experimental results show that PuriGAN is able to deliver the best performance over comparable baselines on both tasks.
☆ Learning End-to-End Channel Coding with Diffusion Models SC
It is a known problem that deep-learning-based end-to-end (E2E) channel coding systems depend on a known and differentiable channel model, due to the learning process and based on the gradient-descent optimization methods. This places the challenge to approximate or generate the channel or its derivative from samples generated by pilot signaling in real-world scenarios. Currently, there are two prevalent methods to solve this problem. One is to generate the channel via a generative adversarial network (GAN), and the other is to, in essence, approximate the gradient via reinforcement learning methods. Other methods include using score-based methods, variational autoencoders, or mutual-information-based methods. In this paper, we focus on generative models and, in particular, on a new promising method called diffusion models, which have shown a higher quality of generation in image-based tasks. We will show that diffusion models can be used in wireless E2E scenarios and that they work as good as Wasserstein GANs while having a more stable training procedure and a better generalization ability in testing.
comment: 6 pages, WSA/SCC 2023
☆ GTV: Generating Tabular Data via Vertical Federated Learning
Generative Adversarial Networks (GANs) have achieved state-of-the-art results in tabular data synthesis, under the presumption of direct accessible training data. Vertical Federated Learning (VFL) is a paradigm which allows to distributedly train machine learning model with clients possessing unique features pertaining to the same individuals, where the tabular data learning is the primary use case. However, it is unknown if tabular GANs can be learned in VFL. Demand for secure data transfer among clients and GAN during training and data synthesizing poses extra challenge. Conditional vector for tabular GANs is a valuable tool to control specific features of generated data. But it contains sensitive information from real data - risking privacy guarantees. In this paper, we propose GTV, a VFL framework for tabular GANs, whose key components are generator, discriminator and the conditional vector. GTV proposes an unique distributed training architecture for generator and discriminator to access training data in a privacy-preserving manner. To accommodate conditional vector into training without privacy leakage, GTV designs a mechanism training-with-shuffling to ensure that no party can reconstruct training data with conditional vector. We evaluate the effectiveness of GTV in terms of synthetic data quality, and overall training scalability. Results show that GTV can consistently generate high-fidelity synthetic tabular data of comparable quality to that generated by centralized GAN algorithm. The difference on machine learning utility can be as low as to 2.7%, even under extremely imbalanced data distributions across clients and different number of clients.
☆ Where and How to Improve Graph-based Spatio-temporal Predictors
This paper introduces a novel residual correlation analysis, called AZ-analysis, to assess the optimality of spatio-temporal predictive models. The proposed AZ-analysis constitutes a valuable asset for discovering and highlighting those space-time regions where the model can be improved with respect to performance. The AZ-analysis operates under very mild assumptions and is based on a spatio-temporal graph that encodes serial and functional dependencies in the data; asymptotically distribution-free summary statistics identify existing residual correlation in space and time regions, hence localizing time frames and/or communities of sensors, where the predictor can be improved.
Better Training of GFlowNets with Local Credit and Incomplete Trajectories
Generative Flow Networks or GFlowNets are related to Monte-Carlo Markov chain methods (as they sample from a distribution specified by an energy function), reinforcement learning (as they learn a policy to sample composed objects through a sequence of steps), generative models (as they learn to represent and sample from a distribution) and amortized variational methods (as they can be used to learn to approximate and sample from an otherwise intractable posterior, given a prior and a likelihood). They are trained to generate an object $x$ through a sequence of steps with probability proportional to some reward function $R(x)$ (or $\exp(-\mathcal{E}(x))$ with $\mathcal{E}(x)$ denoting the energy function), given at the end of the generative trajectory. Like for other RL settings where the reward is only given at the end, the efficiency of training and credit assignment may suffer when those trajectories are longer. With previous GFlowNet work, no learning was possible from incomplete trajectories (lacking a terminal state and the computation of the associated reward). In this paper, we consider the case where the energy function can be applied not just to terminal states but also to intermediate states. This is for example achieved when the energy function is additive, with terms available along the trajectory. We show how to reparameterize the GFlowNet state flow function to take advantage of the partial reward already accrued at each state. This enables a training objective that can be applied to update parameters even with incomplete trajectories. Even when complete trajectories are available, being able to obtain more localized credit and gradients is found to speed up training convergence, as demonstrated across many simulations.
☆ Improving the Timing Resolution of Positron Emission Tomography Detectors using Boosted Learning -- A Residual Physics Approach
Artificial intelligence is finding its way into medical imaging, usually focusing on image reconstruction or enhancing analytical reconstructed images. However, optimizations along the complete processing chain, from detecting signals to computing data, enable significant improvements. Thus, we present an approach toward detector optimization using boosted learning by exploiting the concept of residual physics. In our work, we improve the coincidence time resolution (CTR) of positron emission tomography (PET) detectors. PET enables imaging of metabolic processes by detecting {\gamma}-photons with scintillation detectors. Current research exploits light-sharing detectors, where the scintillation light is distributed over and digitized by an array of readout channels. While these detectors demonstrate excellent performance parameters, e.g., regarding spatial resolution, extracting precise timing information for time-of-flight (TOF) becomes more challenging due to deteriorating effects called time skews. Conventional correction methods mainly rely on analytical formulations, theoretically capable of covering all time skew effects, e.g., caused by signal runtimes or physical effects. However, additional effects are involved for light-sharing detectors, so finding suitable analytical formulations can become arbitrarily complicated. The residual physics-based strategy uses gradient tree boosting (GTB) and a physics-informed data generation mimicking an actual imaging process by shifting a radiation source. We used clinically relevant detectors with a height of 19 mm, coupled to digital photosensor arrays. All trained models improved the CTR significantly. Using the best model, we achieved CTRs down to 198 ps (185 ps) for energies ranging from 300 keV to 700 keV (450 keV to 550 keV).
☆ Two-Stage Constrained Actor-Critic fo Short Video Recommendation
The wide popularity of short videos on social media poses new opportunities and challenges to optimize recommender systems on the video-sharing platforms. Users sequentially interact with the system and provide complex and multi-faceted responses, including watch time and various types of interactions with multiple videos. One the one hand, the platforms aims at optimizing the users' cumulative watch time (main goal) in long term, which can be effectively optimized by Reinforcement Learning. On the other hand, the platforms also needs to satisfy the constraint of accommodating the responses of multiple user interactions (auxiliary goals) such like, follow, share etc. In this paper, we formulate the problem of short video recommendation as a Constrained Markov Decision Process (CMDP). We find that traditional constrained reinforcement learning algorithms can not work well in this setting. We propose a novel two-stage constrained actor-critic method: At stage one, we learn individual policies to optimize each auxiliary signal. At stage two, we learn a policy to (i) optimize the main signal and (ii) stay close to policies learned at the first stage, which effectively guarantees the performance of this main policy on the auxiliaries. Through extensive offline evaluations, we demonstrate effectiveness of our method over alternatives in both optimizing the main goal as well as balancing the others. We further show the advantage of our method in live experiments of short video recommendations, where it significantly outperforms other baselines in terms of both watch time and interactions. Our approach has been fully launched in the production system to optimize user experiences on the platform.
comment: arXiv admin note: substantial text overlap with arXiv:2205.13248
☆ Revisiting Personalized Federated Learning: Robustness Against Backdoor Attacks
In this work, besides improving prediction accuracy, we study whether personalization could bring robustness benefits to backdoor attacks. We conduct the first study of backdoor attacks in the pFL framework, testing 4 widely used backdoor attacks against 6 pFL methods on benchmark datasets FEMNIST and CIFAR-10, a total of 600 experiments. The study shows that pFL methods with partial model-sharing can significantly boost robustness against backdoor attacks. In contrast, pFL methods with full model-sharing do not show robustness. To analyze the reasons for varying robustness performances, we provide comprehensive ablation studies on different pFL methods. Based on our findings, we further propose a lightweight defense method, Simple-Tuning, which empirically improves defense performance against backdoor attacks. We believe that our work could provide both guidance for pFL application in terms of its robustness and offer valuable insights to design more robust FL methods in the future.
☆ Show me your NFT and I tell you how it will perform: Multimodal representation learning for NFT selling price prediction
Non-Fungible Tokens (NFTs) represent deeds of ownership, based on blockchain technologies and smart contracts, of unique crypto assets on digital art forms (e.g., artworks or collectibles). In the spotlight after skyrocketing in 2021, NFTs have attracted the attention of crypto enthusiasts and investors intent on placing promising investments in this profitable market. However, the NFT financial performance prediction has not been widely explored to date. In this work, we address the above problem based on the hypothesis that NFT images and their textual descriptions are essential proxies to predict the NFT selling prices. To this purpose, we propose MERLIN, a novel multimodal deep learning framework designed to train Transformer-based language and visual models, along with graph neural network models, on collections of NFTs' images and texts. A key aspect in MERLIN is its independence on financial features, as it exploits only the primary data a user interested in NFT trading would like to deal with, i.e., NFT images and textual descriptions. By learning dense representations of such data, a price-category classification task is performed by MERLIN models, which can also be tuned according to user preferences in the inference phase to mimic different risk-return investment profiles. Experimental evaluation on a publicly available dataset has shown that MERLIN models achieve significant performances according to several financial assessment criteria, fostering profitable investments, and also beating baseline machine-learning classifiers based on financial features.
comment: Accepted paper at The ACM Web Conference 2023, April 30--May 04, 2023, Austin, Texas, USA
☆ Mind the Gap: Offline Policy Optimization for Imperfect Rewards ICLR2023
Reward function is essential in reinforcement learning (RL), serving as the guiding signal to incentivize agents to solve given tasks, however, is also notoriously difficult to design. In many cases, only imperfect rewards are available, which inflicts substantial performance loss for RL agents. In this study, we propose a unified offline policy optimization approach, \textit{RGM (Reward Gap Minimization)}, which can smartly handle diverse types of imperfect rewards. RGM is formulated as a bi-level optimization problem: the upper layer optimizes a reward correction term that performs visitation distribution matching w.r.t. some expert data; the lower layer solves a pessimistic RL problem with the corrected rewards. By exploiting the duality of the lower layer, we derive a tractable algorithm that enables sampled-based learning without any online interactions. Comprehensive experiments demonstrate that RGM achieves superior performance to existing methods under diverse settings of imperfect rewards. Further, RGM can effectively correct wrong or inconsistent rewards against expert preference and retrieve useful information from biased rewards.
comment: Accept by ICLR2023. The first two authors contributed equally
☆ From slides (through tiles) to pixels: an explainability framework for weakly supervised models in pre-clinical pathology
In pre-clinical pathology, there is a paradox between the abundance of raw data (whole slide images from many organs of many individual animals) and the lack of pixel-level slide annotations done by pathologists. Due to time constraints and requirements from regulatory authorities, diagnoses are instead stored as slide labels. Weakly supervised training is designed to take advantage of those data, and the trained models can be used by pathologists to rank slides by their probability of containing a given lesion of interest. In this work, we propose a novel contextualized eXplainable AI (XAI) framework and its application to deep learning models trained on Whole Slide Images (WSIs) in Digital Pathology. Specifically, we apply our methods to a multi-instance-learning (MIL) model, which is trained solely on slide-level labels, without the need for pixel-level annotations. We validate quantitatively our methods by quantifying the agreements of our explanations' heatmaps with pathologists' annotations, as well as with predictions from a segmentation model trained on such annotations. We demonstrate the stability of the explanations with respect to input shifts, and the fidelity with respect to increased model performance. We quantitatively evaluate the correlation between available pixel-wise annotations and explainability heatmaps. We show that the explanations on important tiles of the whole slide correlate with tissue changes between healthy regions and lesions, but do not exactly behave like a human annotator. This result is coherent with the model training strategy.
comment: 18 pages, 9 figures
☆ Structure-informed Language Models Are Protein Designers
This paper demonstrates that language models are strong structure-based protein designers. We present LM-Design, a generic approach to reprogramming sequence-based protein language models (pLMs), that have learned massive sequential evolutionary knowledge from the universe of natural protein sequences, to acquire an immediate capability to design preferable protein sequences for given folds. We conduct a structural surgery on pLMs, where a lightweight structural adapter is implanted into pLMs and endows it with structural awareness. During inference, iterative refinement is performed to effectively optimize the generated protein sequences. Experiments show that our approach outperforms the state-of-the-art methods by a large margin, leading to up to 4% to 12% accuracy gains in sequence recovery (e.g., 55.65% and 56.63% on CATH 4.2 and 4.3 single-chain benchmarks, and >60% when designing protein complexes). We provide extensive and in-depth analyses, which verify that LM-Design can (1) indeed leverage both structural and sequential knowledge to accurately handle structurally non-deterministic regions, (2) benefit from scaling data and model size, and (3) generalize to other proteins (e.g., antibodies and de novo proteins)
comment: 10 pages
Blockwise Self-Supervised Learning at Scale
Current state-of-the-art deep networks are all powered by backpropagation. In this paper, we explore alternatives to full backpropagation in the form of blockwise learning rules, leveraging the latest developments in self-supervised learning. We show that a blockwise pretraining procedure consisting of training independently the 4 main blocks of layers of a ResNet-50 with Barlow Twins' loss function at each block performs almost as well as end-to-end backpropagation on ImageNet: a linear probe trained on top of our blockwise pretrained model obtains a top-1 classification accuracy of 70.48%, only 1.1% below the accuracy of an end-to-end pretrained network (71.57% accuracy). We perform extensive experiments to understand the impact of different components within our method and explore a variety of adaptations of self-supervised learning to the blockwise paradigm, building an exhaustive understanding of the critical avenues for scaling local learning rules to large networks, with implications ranging from hardware design to neuroscience.
☆ Convergence Analysis of Split Learning on Non-IID Data
Split Learning (SL) is one promising variant of Federated Learning (FL), where the AI model is split and trained at the clients and the server collaboratively. By offloading the computation-intensive portions to the server, SL enables efficient model training on resource-constrained clients. Despite its booming applications, SL still lacks rigorous convergence analysis on non-IID data, which is critical for hyperparameter selection. In this paper, we first prove that SL exhibits an $\mathcal{O}(1/\sqrt{R})$ convergence rate for non-convex objectives on non-IID data, where $R$ is the number of total training rounds. The derived convergence results can facilitate understanding the effect of some crucial factors in SL (e.g., data heterogeneity and synchronization interval). Furthermore, comparing with the convergence result of FL, we show that the guarantee of SL is worse than FL in terms of training rounds on non-IID data. The experimental results verify our theory. More findings on the comparison between FL and SL in cross-device settings are also reported.
☆ Beyond the Universal Law of Robustness: Sharper Laws for Random Features and Neural Tangent Kernels
Machine learning models are vulnerable to adversarial perturbations, and a thought-provoking paper by Bubeck and Sellke has analyzed this phenomenon through the lens of over-parameterization: interpolating smoothly the data requires significantly more parameters than simply memorizing it. However, this "universal" law provides only a necessary condition for robustness, and it is unable to discriminate between models. In this paper, we address these gaps by focusing on empirical risk minimization in two prototypical settings, namely, random features and the neural tangent kernel (NTK). We prove that, for random features, the model is not robust for any degree of over-parameterization, even when the necessary condition coming from the universal law of robustness is satisfied. In contrast, for even activations, the NTK model meets the universal lower bound, and it is robust as soon as the necessary condition on over-parameterization is fulfilled. This also addresses a conjecture in prior work by Bubeck, Li and Nagaraj. Our analysis decouples the effect of the kernel of the model from an "interaction matrix", which describes the interaction with the test data and captures the effect of the activation. Our theoretical results are corroborated by numerical evidence on both synthetic and standard datasets (MNIST, CIFAR-10).
☆ Private, fair and accurate: Training large-scale, privacy-preserving AI models in radiology
Artificial intelligence (AI) models are increasingly used in the medical domain. However, as medical data is highly sensitive, special precautions to ensure the protection of said data are required. The gold standard for privacy preservation is the introduction of differential privacy (DP) to model training. However, prior work has shown that DP has negative implications on model accuracy and fairness. Therefore, the purpose of this study is to demonstrate that the privacy-preserving training of AI models for chest radiograph diagnosis is possible with high accuracy and fairness compared to non-private training. N=193,311 high quality clinical chest radiographs were retrospectively collected and manually labeled by experienced radiologists, who assigned one or more of the following diagnoses: cardiomegaly, congestion, pleural effusion, pneumonic infiltration and atelectasis, to each side (where applicable). The non-private AI models were compared with privacy-preserving (DP) models with respect to privacy-utility trade-offs (measured as area under the receiver-operator-characteristic curve (AUROC)), and privacy-fairness trade-offs (measured as Pearson-R or Statistical Parity Difference). The non-private AI model achieved an average AUROC score of 0.90 over all labels, whereas the DP AI model with a privacy budget of epsilon=7.89 resulted in an AUROC of 0.87, i.e., a mere 2.6% performance decrease compared to non-private training. The privacy-preserving training of diagnostic AI models can achieve high performance with a small penalty on model accuracy and does not amplify discrimination against age, sex or co-morbidity. We thus encourage practitioners to integrate state-of-the-art privacy-preserving techniques into medical AI model development.
comment: 3 tables, 5 figures, 11 supplementary materials
☆ A Feature Selection Method for Driver Stress Detection Using Heart Rate Variability and Breathing Rate
Driver stress is a major cause of car accidents and death worldwide. Furthermore, persistent stress is a health problem, contributing to hypertension and other diseases of the cardiovascular system. Stress has a measurable impact on heart and breathing rates and stress levels can be inferred from such measurements. Galvanic skin response is a common test to measure the perspiration caused by both physiological and psychological stress, as well as extreme emotions. In this paper, galvanic skin response is used to estimate the ground truth stress levels. A feature selection technique based on the minimal redundancy-maximal relevance method is then applied to multiple heart rate variability and breathing rate metrics to identify a novel and optimal combination for use in detecting stress. The support vector machine algorithm with a radial basis function kernel was used along with these features to reliably predict stress. The proposed method has achieved a high level of accuracy on the target dataset.
comment: In Proceedings of the 15th International Conference on Machine Vision (ICMV), Romem Italy, 18-20 November 2022. arXiv admin note: text overlap with arXiv:2206.03222
☆ SCCAM: Supervised Contrastive Convolutional Attention Mechanism for Ante-hoc Interpretable Fault Diagnosis with Limited Fault Samples
In real industrial processes, fault diagnosis methods are required to learn from limited fault samples since the procedures are mainly under normal conditions and the faults rarely occur. Although attention mechanisms have become popular in the field of fault diagnosis, the existing attention-based methods are still unsatisfying for the above practical applications. First, pure attention-based architectures like transformers need a large number of fault samples to offset the lack of inductive biases thus performing poorly under limited fault samples. Moreover, the poor fault classification dilemma further leads to the failure of the existing attention-based methods to identify the root causes. To address the aforementioned issues, we innovatively propose a supervised contrastive convolutional attention mechanism (SCCAM) with ante-hoc interpretability, which solves the root cause analysis problem under limited fault samples for the first time. The proposed SCCAM method is tested on a continuous stirred tank heater and the Tennessee Eastman industrial process benchmark. Three common fault diagnosis scenarios are covered, including a balanced scenario for additional verification and two scenarios with limited fault samples (i.e., imbalanced scenario and long-tail scenario). The comprehensive results demonstrate that the proposed SCCAM method can achieve better performance compared with the state-of-the-art methods on fault classification and root cause analysis.
☆ A Novel Fuzzy Bi-Clustering Algorithm with AFS for Identification of Co-Regulated Genes
The identification of co-regulated genes and their transcription-factor binding sites (TFBS) are the key steps toward understanding transcription regulation. In addition to effective laboratory assays, various bi-clustering algorithms for detection of the co-expressed genes have been developed. Bi-clustering methods are used to discover subgroups of genes with similar expression patterns under to-be-identified subsets of experimental conditions when applied to gene expression data. By building two fuzzy partition matrices of the gene expression data with the Axiomatic Fuzzy Set (AFS) theory, this paper proposes a novel fuzzy bi-clustering algorithm for identification of co-regulated genes. Specifically, the gene expression data is transformed into two fuzzy partition matrices via sub-preference relations theory of AFS at first. One of the matrices is considering the genes as the universe and the conditions as the concept, the other one is considering the genes as the concept and the conditions as the universe. The identification of the co-regulated genes (bi-clusters) is carried out on the two partition matrices at the same time. Then, a novel fuzzy-based similarity criterion is defined based on the partition matrixes, and a cyclic optimization algorithm is designed to discover the significant bi-clusters at expression level. The above procedures guarantee that the generated bi-clusters have more significant expression values than that of extracted by the traditional bi-clustering methods. Finally, the performance of the proposed method is evaluated with the performance of the three well-known bi-clustering algorithms on publicly available real microarray datasets. The experimental results are in agreement with the theoretical analysis and show that the proposed algorithm can effectively detect the co-regulated genes without any prior knowledge of the gene expression data.
☆ Deep Reinforcement Learning for Cyber System Defense under Dynamic Adversarial Uncertainties
Development of autonomous cyber system defense strategies and action recommendations in the real-world is challenging, and includes characterizing system state uncertainties and attack-defense dynamics. We propose a data-driven deep reinforcement learning (DRL) framework to learn proactive, context-aware, defense countermeasures that dynamically adapt to evolving adversarial behaviors while minimizing loss of cyber system operations. A dynamic defense optimization problem is formulated with multiple protective postures against different types of adversaries with varying levels of skill and persistence. A custom simulation environment was developed and experiments were devised to systematically evaluate the performance of four model-free DRL algorithms against realistic, multi-stage attack sequences. Our results suggest the efficacy of DRL algorithms for proactive cyber defense under multi-stage attack profiles and system uncertainties.
☆ Learning to Decouple Complex Systems
A complex system with cluttered observations may be a coupled mixture of multiple simple sub-systems corresponding to latent entities. Such sub-systems may hold distinct dynamics in the continuous-time domain; therein, complicated interactions between sub-systems also evolve over time. This setting is fairly common in the real world but has been less considered. In this paper, we propose a sequential learning approach under this setting by decoupling a complex system for handling irregularly sampled and cluttered sequential observations. Such decoupling brings about not only subsystems describing the dynamics of each latent entity but also a meta-system capturing the interaction between entities over time. Specifically, we argue that the meta-system evolving within a simplex is governed by projected differential equations (ProjDEs). We further analyze and provide neural-friendly projection operators in the context of Bregman divergence. Experimental results on synthetic and real-world datasets show the advantages of our approach when facing complex and cluttered sequential data compared to the state-of-the-art.
☆ ResMem: Learn what you can and memorize the rest
The impressive generalization performance of modern neural networks is attributed in part to their ability to implicitly memorize complex training patterns. Inspired by this, we explore a novel mechanism to improve model generalization via explicit memorization. Specifically, we propose the residual-memorization (ResMem) algorithm, a new method that augments an existing prediction model (e.g. a neural network) by fitting the model's residuals with a $k$-nearest neighbor based regressor. The final prediction is then the sum of the original model and the fitted residual regressor. By construction, ResMem can explicitly memorize the training labels. Empirically, we show that ResMem consistently improves the test set generalization of the original prediction model across various standard vision and natural language processing benchmarks. Theoretically, we formulate a stylized linear regression problem and rigorously show that ResMem results in a more favorable test risk over the base predictor.
☆ An Operational Perspective to Fairness Interventions: Where and How to Intervene
As AI-based decision systems proliferate, their successful operationalization requires balancing multiple desiderata: predictive performance, disparity across groups, safeguarding sensitive group attributes (e.g., race), and engineering cost. We present a holistic framework for evaluating and contextualizing fairness interventions with respect to the above desiderata. The two key points of practical consideration are where (pre-, in-, post-processing) and how (in what way the sensitive group data is used) the intervention is introduced. We demonstrate our framework using a thorough benchmarking study on predictive parity; we study close to 400 methodological variations across two major model types (XGBoost vs. Neural Net) and ten datasets. Methodological insights derived from our empirical study inform the practical design of ML workflow with fairness as a central concern. We find predictive parity is difficult to achieve without using group data, and despite requiring group data during model training (but not inference), distributionally robust methods provide significant Pareto improvement. Moreover, a plain XGBoost model often Pareto-dominates neural networks with fairness interventions, highlighting the importance of model inductive bias.
☆ Robust Camera Pose Refinement for Multi-Resolution Hash Encoding
Multi-resolution hash encoding has recently been proposed to reduce the computational cost of neural renderings, such as NeRF. This method requires accurate camera poses for the neural renderings of given scenes. However, contrary to previous methods jointly optimizing camera poses and 3D scenes, the naive gradient-based camera pose refinement method using multi-resolution hash encoding severely deteriorates performance. We propose a joint optimization algorithm to calibrate the camera pose and learn a geometric representation using efficient multi-resolution hash encoding. Showing that the oscillating gradient flows of hash encoding interfere with the registration of camera poses, our method addresses the issue by utilizing smooth interpolation weighting to stabilize the gradient oscillation for the ray samplings across hash grids. Moreover, the curriculum training procedure helps to learn the level-wise hash encoding, further increasing the pose refinement. Experiments on the novel-view synthesis datasets validate that our learning frameworks achieve state-of-the-art performance and rapid convergence of neural rendering, even when initial camera poses are unknown.
☆ Uniform tensor clustering by jointly exploring sample affinities of various orders
Conventional clustering methods based on pairwise affinity usually suffer from the concentration effect while processing huge dimensional features yet low sample sizes data, resulting in inaccuracy to encode the sample proximity and suboptimal performance in clustering. To address this issue, we propose a unified tensor clustering method (UTC) that characterizes sample proximity using multiple samples' affinity, thereby supplementing rich spatial sample distributions to boost clustering. Specifically, we find that the triadic tensor affinity can be constructed via the Khari-Rao product of two affinity matrices. Furthermore, our early work shows that the fourth-order tensor affinity is defined by the Kronecker product. Therefore, we utilize arithmetical products, Khatri-Rao and Kronecker products, to mathematically integrate different orders of affinity into a unified tensor clustering framework. Thus, the UTC jointly learns a joint low-dimensional embedding to combine various orders. Finally, a numerical scheme is designed to solve the problem. Experiments on synthetic datasets and real-world datasets demonstrate that 1) the usage of high-order tensor affinity could provide a supplementary characterization of sample proximity to the popular affinity matrix; 2) the proposed method of UTC is affirmed to enhance clustering by exploiting different order affinities when processing high-dimensional data.
☆ DynaMIX: Resource Optimization for DNN-Based Real-Time Applications on a Multi-Tasking System
As deep neural networks (DNNs) prove their importance and feasibility, more and more DNN-based apps, such as detection and classification of objects, have been developed and deployed on autonomous vehicles (AVs). To meet their growing expectations and requirements, AVs should "optimize" use of their limited onboard computing resources for multiple concurrent in-vehicle apps while satisfying their timing requirements (especially for safety). That is, real-time AV apps should share the limited on-board resources with other concurrent apps without missing their deadlines dictated by the frame rate of a camera that generates and provides input images to the apps. However, most, if not all, of existing DNN solutions focus on enhancing the concurrency of their specific hardware without dynamically optimizing/modifying the DNN apps' resource requirements, subject to the number of running apps, owing to their high computational cost. To mitigate this limitation, we propose DynaMIX (Dynamic MIXed-precision model construction), which optimizes the resource requirement of concurrent apps and aims to maximize execution accuracy. To realize a real-time resource optimization, we formulate an optimization problem using app performance profiles to consider both the accuracy and worst-case latency of each app. We also propose dynamic model reconfiguration by lazy loading only the selected layers at runtime to reduce the overhead of loading the entire model. DynaMIX is evaluated in terms of constraint satisfaction and inference accuracy for a multi-tasking system and compared against state-of-the-art solutions, demonstrating its effectiveness and feasibility under various environmental/operating conditions.
comment: 13 pages, 9 figures, 5 tables
☆ Deep Reinforcement Learning for Online Error Detection in Cyber-Physical Systems
Reliability is one of the major design criteria in Cyber-Physical Systems (CPSs). This is because of the existence of some critical applications in CPSs and their failure is catastrophic. Therefore, employing strong error detection and correction mechanisms in CPSs is inevitable. CPSs are composed of a variety of units, including sensors, networks, and microcontrollers. Each of these units is probable to be in a faulty state at any time and the occurred fault can result in erroneous output. The fault may cause the units of CPS to malfunction and eventually crash. Traditional fault-tolerant approaches include redundancy time, hardware, information, and/or software. However, these approaches impose significant overheads besides their low error coverage, which limits their applicability. In addition, the interval between error occurrence and detection is too long in these approaches. In this paper, based on Deep Reinforcement Learning (DRL), a new error detection approach is proposed that not only detects errors with high accuracy but also can perform error detection at the moment due to very low inference time. The proposed approach can categorize different types of errors from normal data and predict whether the system will fail. The evaluation results illustrate that the proposed approach has improved more than 2x in terms of accuracy and more than 5x in terms of inference time compared to other approaches.
☆ Causal Inference Based Single-branch Ensemble Trees For Uplift Modeling
In this manuscript (ms), we propose causal inference based single-branch ensemble trees for uplift modeling, namely CIET. Different from standard classification methods for predictive probability modeling, CIET aims to achieve the change in the predictive probability of outcome caused by an action or a treatment. According to our CIET, two partition criteria are specifically designed to maximize the difference in outcome distribution between the treatment and control groups. Next, a novel single-branch tree is built by taking a top-down node partition approach, and the remaining samples are censored since they are not covered by the upper node partition logic. Repeating the tree-building process on the censored data, single-branch ensemble trees with a set of inference rules are thus formed. Moreover, CIET is experimentally demonstrated to outperform previous approaches for uplift modeling in terms of both area under uplift curve (AUUC) and Qini coefficient significantly. At present, CIET has already been applied to online personal loans in a national financial holdings group in China. CIET will also be of use to analysts applying machine learning techniques to causal inference in broader business domains such as web advertising, medicine and economics.
☆ Machine Learning for UAV Propeller Fault Detection based on a Hybrid Data Generation Model
This paper describes the development of an on-board data-driven system that can monitor and localize the fault in a quadrotor unmanned aerial vehicle (UAV) and at the same time, evaluate the degree of damage of the fault under real scenarios. To achieve offline training data generation, a hybrid approach is proposed for the development of a virtual data-generative model using a combination of data-driven models as well as well-established dynamic models that describe the kinematics of the UAV. To effectively represent the drop in performance of a faulty propeller, a variation of the deep neural network, a LSTM network is proposed. With the RPM of the propeller as input and based on the fault condition of the propeller, the proposed propeller model estimates the resultant torque and thrust. Then, flight datasets of the UAV under various fault scenarios are generated via simulation using the developed data-generative model. Lastly, a fault classifier using a CNN model is proposed to identify as well as evaluate the degree of damage to the damaged propeller. The scope of this paper focuses on the identification of faulty propellers and classification of the fault level for quadrotor UAVs using RPM as well as flight data. Doing so allows for early minor fault detection to prevent serious faults from occurring if the fault is left unrepaired. To further validate the workability of this approach outside of simulation, a real-flight test is conducted indoors. The real flight data is collected and a simulation to real sim-real test is conducted. Due to the imperfections in the build of our experimental UAV, a slight calibration approach to our simulation model is further proposed and the experimental results obtained show that our trained model can identify the location of propeller fault as well as the degree/type of damage. Currently, the diagnosis accuracy on the testing set is over 80%.
☆ Vertical Federated Learning: Taxonomies, Threats, and Prospects
Federated learning (FL) is the most popular distributed machine learning technique. FL allows machine-learning models to be trained without acquiring raw data to a single point for processing. Instead, local models are trained with local data; the models are then shared and combined. This approach preserves data privacy as locally trained models are shared instead of the raw data themselves. Broadly, FL can be divided into horizontal federated learning (HFL) and vertical federated learning (VFL). For the former, different parties hold different samples over the same set of features; for the latter, different parties hold different feature data belonging to the same set of samples. In a number of practical scenarios, VFL is more relevant than HFL as different companies (e.g., bank and retailer) hold different features (e.g., credit history and shopping history) for the same set of customers. Although VFL is an emerging area of research, it is not well-established compared to HFL. Besides, VFL-related studies are dispersed, and their connections are not intuitive. Thus, this survey aims to bring these VFL-related studies to one place. Firstly, we classify existing VFL structures and algorithms. Secondly, we present the threats from security and privacy perspectives to VFL. Thirdly, for the benefit of future researchers, we discussed the challenges and prospects of VFL in detail.
☆ Group Fairness in Non-monotone Submodular Maximization
Maximizing a submodular function has a wide range of applications in machine learning and data mining. One such application is data summarization whose goal is to select a small set of representative and diverse data items from a large dataset. However, data items might have sensitive attributes such as race or gender, in this setting, it is important to design \emph{fairness-aware} algorithms to mitigate potential algorithmic bias that may cause over- or under- representation of particular groups. Motivated by that, we propose and study the classic non-monotone submodular maximization problem subject to novel group fairness constraints. Our goal is to select a set of items that maximizes a non-monotone submodular function, while ensuring that the number of selected items from each group is proportionate to its size, to the extent specified by the decision maker. We develop the first constant-factor approximation algorithms for this problem. We also extend the basic model to incorporate an additional global size constraint on the total number of selected items.
☆ Optimality of Thompson Sampling with Noninformative Priors for Pareto Bandits
In the stochastic multi-armed bandit problem, a randomized probability matching policy called Thompson sampling (TS) has shown excellent performance in various reward models. In addition to the empirical performance, TS has been shown to achieve asymptotic problem-dependent lower bounds in several models. However, its optimality has been mainly addressed under light-tailed or one-parameter models that belong to exponential families. In this paper, we consider the optimality of TS for the Pareto model that has a heavy tail and is parameterized by two unknown parameters. Specifically, we discuss the optimality of TS with probability matching priors that include the Jeffreys prior and the reference priors. We first prove that TS with certain probability matching priors can achieve the optimal regret bound. Then, we show the suboptimality of TS with other priors, including the Jeffreys and the reference priors. Nevertheless, we find that TS with the Jeffreys and reference priors can achieve the asymptotic lower bound if one uses a truncation procedure. These results suggest carefully choosing noninformative priors to avoid suboptimality and show the effectiveness of truncation procedures in TS-based policies.
comment: 49 pages, a preprint
☆ Multiplier Bootstrap-based Exploration
Despite the great interest in the bandit problem, designing efficient algorithms for complex models remains challenging, as there is typically no analytical way to quantify uncertainty. In this paper, we propose Multiplier Bootstrap-based Exploration (MBE), a novel exploration strategy that is applicable to any reward model amenable to weighted loss minimization. We prove both instance-dependent and instance-independent rate-optimal regret bounds for MBE in sub-Gaussian multi-armed bandits. With extensive simulation and real data experiments, we show the generality and adaptivity of MBE.
☆ A Lipschitz Bandits Approach for Continuous Hyperparameter Optimization
One of the most critical problems in machine learning is HyperParameter Optimization (HPO), since choice of hyperparameters has a significant impact on final model performance. Although there are many HPO algorithms, they either have no theoretical guarantees or require strong assumptions. To this end, we introduce BLiE -- a Lipschitz-bandit-based algorithm for HPO that only assumes Lipschitz continuity of the objective function. BLiE exploits the landscape of the objective function to adaptively search over the hyperparameter space. Theoretically, we show that $(i)$ BLiE finds an $\epsilon$-optimal hyperparameter with $O \left( \frac{1}{\epsilon} \right)^{d_z + \beta}$ total budgets, where $d_z$ and $\beta$ are problem intrinsic; $(ii)$ BLiE is highly parallelizable. Empirically, we demonstrate that BLiE outperforms the state-of-the-art HPO algorithms on benchmark tasks. We also apply BLiE to search for noise schedule of diffusion models. Comparison with the default schedule shows that BLiE schedule greatly improves the sampling speed.
☆ DCM: Deep energy method based on the principle of minimum complementary energy
The principle of minimum potential and complementary energy are the most important variational principles in solid mechanics. The deep energy method (DEM), which has received much attention, is based on the principle of minimum potential energy and lacks the important form of minimum complementary energy. Thus, we propose the deep energy method based on the principle of minimum complementary energy (DCM). The output function of DCM is the stress function that naturally satisfies the equilibrium equation. We extend the proposed DCM algorithm (DCM-P), adding the terms that naturally satisfy the biharmonic equation in the Airy stress function. We combine operator learning with physical equations and propose a deep complementary energy operator method (DCM-O), including branch net, trunk net, basis net, and particular net. DCM-O first combines existing high-fidelity numerical results to train DCM-O through data. Then the complementary energy is used to train the branch net and trunk net in DCM-O. To analyze DCM performance, we present the numerical result of the most common stress functions, the Prandtl and Airy stress function. The proposed method DCM is used to model the representative mechanical problems with the different types of boundary conditions. We compare DCM with the existing PINNs and DEM algorithms. The result shows the advantage of the proposed DCM is suitable for dealing with problems of dominated displacement boundary conditions, which is reflected in theory and our numerical experiments. DCM-P and DCM-O improve the accuracy of DCM and the speed of calculation convergence. DCM is an essential supplementary energy form of the deep energy method. We believe that operator learning based on the energy method can balance data and physical equations well, giving computational mechanics broad research prospects.
comment: 46 pages, 23 figures
☆ Using natural language processing and structured medical data to phenotype patients hospitalized due to COVID-19
To identify patients who are hospitalized because of COVID-19 as opposed to those who were admitted for other indications, we compared the performance of different computable phenotype definitions for COVID-19 hospitalizations that use different types of data from the electronic health records (EHR), including structured EHR data elements, provider notes, or a combination of both data types. And conduct a retrospective data analysis utilizing chart review-based validation. Participants are 586 hospitalized individuals who tested positive for SARS-CoV-2 during January 2022. We used natural language processing to incorporate data from provider notes and LASSO regression and Random Forests to fit classification algorithms that incorporated structured EHR data elements, provider notes, or a combination of structured data and provider notes. Results: Based on a chart review, 38% of 586 patients were determined to be hospitalized for reasons other than COVID-19 despite having tested positive for SARS-CoV-2. A classification algorithm that used provider notes had significantly better discrimination than one that used structured EHR data elements (AUROC: 0.894 vs 0.841, p < 0.001), and performed similarly to a model that combined provider notes with structured data elements (AUROC: 0.894 vs 0.893). Assessments of hospital outcome metrics significantly differed based on whether the population included all hospitalized patients who tested positive for SARS-CoV-2 versus those who were determined to have been hospitalized due to COVID-19. This work demonstrates the utility of natural language processing approaches to derive information related to patient hospitalizations in cases where there may be multiple conditions that could serve as the primary indication for hospitalization.
comment: 21 pages, 2 figures, 3 tables, 1 supplemental figure, 2 supplemental tables
☆ Support Recovery in Sparse PCA with Non-Random Missing Data
We analyze a practical algorithm for sparse PCA on incomplete and noisy data under a general non-random sampling scheme. The algorithm is based on a semidefinite relaxation of the $\ell_1$-regularized PCA problem. We provide theoretical justification that under certain conditions, we can recover the support of the sparse leading eigenvector with high probability by obtaining a unique solution. The conditions involve the spectral gap between the largest and second-largest eigenvalues of the true data matrix, the magnitude of the noise, and the structural properties of the observed entries. The concepts of algebraic connectivity and irregularity are used to describe the structural properties of the observed entries. We empirically justify our theorem with synthetic and real data analysis. We also show that our algorithm outperforms several other sparse PCA approaches especially when the observed entries have good structural properties. As a by-product of our analysis, we provide two theorems to handle a deterministic sampling scheme, which can be applied to other matrix-related problems.
comment: arXiv admin note: text overlap with arXiv:2205.15215
☆ Revisiting Intermediate Layer Distillation for Compressing Language Models: An Overfitting Perspective
Knowledge distillation (KD) is a highly promising method for mitigating the computational problems of pre-trained language models (PLMs). Among various KD approaches, Intermediate Layer Distillation (ILD) has been a de facto standard KD method with its performance efficacy in the NLP field. In this paper, we find that existing ILD methods are prone to overfitting to training datasets, although these methods transfer more information than the original KD. Next, we present the simple observations to mitigate the overfitting of ILD: distilling only the last Transformer layer and conducting ILD on supplementary tasks. Based on our two findings, we propose a simple yet effective consistency-regularized ILD (CR-ILD), which prevents the student model from overfitting the training dataset. Substantial experiments on distilling BERT on the GLUE benchmark and several synthetic datasets demonstrate that our proposed ILD method outperforms other KD techniques. Our code is available at https://github.com/jongwooko/CR-ILD.
comment: The 17th Conference of the European Chapter of the Association for Computational Linguistics (Findings)
☆ Example-Based Explainable AI and its Application for Remote Sensing Image Classification
We present a method of explainable artificial intelligence (XAI), "What I Know (WIK)", to provide additional information to verify the reliability of a deep learning model by showing an example of an instance in a training dataset that is similar to the input data to be inferred and demonstrate it in a remote sensing image classification task. One of the expected roles of XAI methods is verifying whether inferences of a trained machine learning model are valid for an application, and it is an important factor that what datasets are used for training the model as well as the model architecture. Our data-centric approach can help determine whether the training dataset is sufficient for each inference by checking the selected example data. If the selected example looks similar to the input data, we can confirm that the model was not trained on a dataset with a feature distribution far from the feature of the input data. With this method, the criteria for selecting an example are not merely data similarity with the input data but also data similarity in the context of the model task. Using a remote sensing image dataset from the Sentinel-2 satellite, the concept was successfully demonstrated with reasonably selected examples. This method can be applied to various machine-learning tasks, including classification and regression.
comment: 10 pages, 4 figures, accepted for publication in International Journal of Applied Earth Observation and Geoinformation
☆ Ordered GNN: Ordering Message Passing to Deal with Heterophily and Over-smoothing ICLR 2023
Most graph neural networks follow the message passing mechanism. However, it faces the over-smoothing problem when multiple times of message passing is applied to a graph, causing indistinguishable node representations and prevents the model to effectively learn dependencies between farther-away nodes. On the other hand, features of neighboring nodes with different labels are likely to be falsely mixed, resulting in the heterophily problem. In this work, we propose to order the messages passing into the node representation, with specific blocks of neurons targeted for message passing within specific hops. This is achieved by aligning the hierarchy of the rooted-tree of a central node with the ordered neurons in its node representation. Experimental results on an extensive set of datasets show that our model can simultaneously achieve the state-of-the-art in both homophily and heterophily settings, without any targeted design. Moreover, its performance maintains pretty well while the model becomes really deep, effectively preventing the over-smoothing problem. Finally, visualizing the gating vectors shows that our model learns to behave differently between homophily and heterophily settings, providing an explainable graph neural model.
comment: Published as a conference paper at ICLR 2023
☆ Multi-channel Autobidding with Budget and ROI Constraints
In digital online advertising, advertisers procure ad impressions simultaneously on multiple platforms, or so-called channels, such as Google Ads, Meta Ads Manager, etc., each of which consists of numerous ad auctions. We study how an advertiser maximizes total conversion (e.g. ad clicks) while satisfying aggregate return-on-investment (ROI) and budget constraints across all channels. In practice, an advertiser does not have control over, and thus cannot globally optimize, which individual ad auctions she participates in for each channel, and instead authorizes a channel to procure impressions on her behalf: the advertiser can only utilize two levers on each channel, namely setting a per-channel budget and per-channel target ROI. In this work, we first analyze the effectiveness of each of these levers for solving the advertiser's global multi-channel problem. We show that when an advertiser only optimizes over per-channel ROIs, her total conversion can be arbitrarily worse than what she could have obtained in the global problem. Further, we show that the advertiser can achieve the global optimal conversion when she only optimizes over per-channel budgets. In light of this finding, under a bandit feedback setting that mimics real-world scenarios where advertisers have limited information on ad auctions in each channels and how channels procure ads, we present an efficient learning algorithm that produces per-channel budgets whose resulting conversion approximates that of the global optimal problem. Finally, we argue that all our results hold for both single-item and multi-item auctions from which channels procure impressions on advertisers' behalf.
☆ Improving Recommendation Relevance by simulating User Interest
Most if not all on-line item-to-item recommendation systems rely on estimation of a distance like measure (rank) of similarity between items. For on-line recommendation systems, time sensitivity of this similarity measure is extremely important. We observe that recommendation "recency" can be straightforwardly and transparently maintained by iterative reduction of ranks of inactive items. The paper briefly summarizes algorithmic developments based on this self-explanatory observation. The basic idea behind this work is patented in a context of online recommendation systems.
☆ LSA-PINN: Linear Boundary Connectivity Loss for Solving PDEs on Complex Geometry
We present a novel loss formulation for efficient learning of complex dynamics from governing physics, typically described by partial differential equations (PDEs), using physics-informed neural networks (PINNs). In our experiments, existing versions of PINNs are seen to learn poorly in many problems, especially for complex geometries, as it becomes increasingly difficult to establish appropriate sampling strategy at the near boundary region. Overly dense sampling can adversely impede training convergence if the local gradient behaviors are too complex to be adequately modelled by PINNs. On the other hand, if the samples are too sparse, existing PINNs tend to overfit the near boundary region, leading to incorrect solution. To prevent such issues, we propose a new Boundary Connectivity (BCXN) loss function which provides linear local structure approximation (LSA) to the gradient behaviors at the boundary for PINN. Our BCXN-loss implicitly imposes local structure during training, thus facilitating fast physics-informed learning across entire problem domains with order of magnitude sparser training samples. This LSA-PINN method shows a few orders of magnitude smaller errors than existing methods in terms of the standard L2-norm metric, while using dramatically fewer training samples and iterations. Our proposed LSA-PINN does not pose any requirement on the differentiable property of the networks, and we demonstrate its benefits and ease of implementation on both multi-layer perceptron and convolutional neural network versions as commonly used in current PINN literature.
comment: 11 pages, 7 figures
☆ Pseudonorm Approachability and Applications to Regret Minimization ALT 2023
Blackwell's celebrated approachability theory provides a general framework for a variety of learning problems, including regret minimization. However, Blackwell's proof and implicit algorithm measure approachability using the $\ell_2$ (Euclidean) distance. We argue that in many applications such as regret minimization, it is more useful to study approachability under other distance metrics, most commonly the $\ell_\infty$-metric. But, the time and space complexity of the algorithms designed for $\ell_\infty$-approachability depend on the dimension of the space of the vectorial payoffs, which is often prohibitively large. Thus, we present a framework for converting high-dimensional $\ell_\infty$-approachability problems to low-dimensional pseudonorm approachability problems, thereby resolving such issues. We first show that the $\ell_\infty$-distance between the average payoff and the approachability set can be equivalently defined as a pseudodistance between a lower-dimensional average vector payoff and a new convex set we define. Next, we develop an algorithmic theory of pseudonorm approachability, analogous to previous work on approachability for $\ell_2$ and other norms, showing that it can be achieved via online linear optimization (OLO) over a convex set given by the Fenchel dual of the unit pseudonorm ball. We then use that to show, modulo mild normalization assumptions, that there exists an $\ell_\infty$-approachability algorithm whose convergence is independent of the dimension of the original vectorial payoff. We further show that that algorithm admits a polynomial-time complexity, assuming that the original $\ell_\infty$-distance can be computed efficiently. We also give an $\ell_\infty$-approachability algorithm whose convergence is logarithmic in that dimension using an FTRL algorithm with a maximum-entropy regularizer.
comment: To appear at ALT 2023
☆ Towards Practical Preferential Bayesian Optimization with Skew Gaussian Processes
We study preferential Bayesian optimization (BO) where reliable feedback is limited to pairwise comparison called duels. An important challenge in preferential BO, which uses the preferential Gaussian process (GP) model to represent flexible preference structure, is that the posterior distribution is a computationally intractable skew GP. The most widely used approach for preferential BO is Gaussian approximation, which ignores the skewness of the true posterior. Alternatively, Markov chain Monte Carlo (MCMC) based preferential BO is also proposed. In this work, we first verify the accuracy of Gaussian approximation, from which we reveal the critical problem that the predictive probability of duels can be inaccurate. This observation motivates us to improve the MCMC-based estimation for skew GP, for which we show the practical efficiency of Gibbs sampling and derive the low variance MC estimator. However, the computational time of MCMC can still be a bottleneck in practice. Towards building a more practical preferential BO, we develop a new method that achieves both high computational efficiency and low sample complexity, and then demonstrate its effectiveness through extensive numerical experiments.
☆ Randomized Gaussian Process Upper Confidence Bound with Tight Bayesian Regret Bounds
Gaussian process upper confidence bound (GP-UCB) is a theoretically promising approach for black-box optimization; however, the confidence parameter $\beta$ is considerably large in the theorem and chosen heuristically in practice. Then, randomized GP-UCB (RGP-UCB) uses a randomized confidence parameter, which follows the Gamma distribution, to mitigate the impact of manually specifying $\beta$. This study first generalizes the regret analysis of RGP-UCB to a wider class of distributions, including the Gamma distribution. Furthermore, we propose improved RGP-UCB (IRGP-UCB) based on a two-parameter exponential distribution, which achieves tight Bayesian regret bounds. IRGP-UCB does not require an increase in the confidence parameter in terms of the number of iterations, which avoids over-exploration in the later iterations. Finally, we demonstrate the effectiveness of IRGP-UCB through extensive experiments.
☆ LazyGNN: Large-Scale Graph Neural Networks via Lazy Propagation
Recent works have demonstrated the benefits of capturing long-distance dependency in graphs by deeper graph neural networks (GNNs). But deeper GNNs suffer from the long-lasting scalability challenge due to the neighborhood explosion problem in large-scale graphs. In this work, we propose to capture long-distance dependency in graphs by shallower models instead of deeper models, which leads to a much more efficient model, LazyGNN, for graph representation learning. Moreover, we demonstrate that LazyGNN is compatible with existing scalable approaches (such as sampling methods) for further accelerations through the development of mini-batch LazyGNN. Comprehensive experiments demonstrate its superior prediction performance and scalability on large-scale benchmarks. LazyGNN also achieves state-of-art performance on the OGB leaderboard.
☆ ANTM: An Aligned Neural Topic Model for Exploring Evolving Topics
As the amount of text data generated by humans and machines increases, the necessity of understanding large corpora and finding a way to extract insights from them is becoming more crucial than ever. Dynamic topic models are effective methods that primarily focus on studying the evolution of topics present in a collection of documents. These models are widely used for understanding trends, exploring public opinion in social networks, or tracking research progress and discoveries in scientific archives. Since topics are defined as clusters of semantically similar documents, it is necessary to observe the changes in the content or themes of these clusters in order to understand how topics evolve as new knowledge is discovered over time. In this paper, we introduce the Aligned Neural Topic Model (ANTM), a dynamic neural topic model that uses document embeddings to compute clusters of semantically similar documents at different periods and to align document clusters to represent their evolution. This alignment procedure preserves the temporal similarity of document clusters over time and captures the semantic change of words characterized by their context within different periods. Experiments on four different datasets show that ANTM outperforms probabilistic dynamic topic models (e.g. DTM, DETM) and significantly improves topic coherence and diversity over other existing dynamic neural topic models (e.g. BERTopic).
☆ Spiking Synaptic Penalty: Appropriate Penalty Term for Energy-Efficient Spiking Neural Networks
Spiking neural networks (SNNs) are energy-efficient neural networks because of their spiking nature. However, as the spike firing rate of SNNs increases, the energy consumption does as well, and thus, the advantage of SNNs diminishes. Here, we tackle this problem by introducing a novel penalty term for the spiking activity into the objective function in the training phase. Our method is designed so as to optimize the energy consumption metric directly without modifying the network architecture. Therefore, the proposed method can reduce the energy consumption more than other methods while maintaining the accuracy. We conducted experiments for image classification tasks, and the results indicate the effectiveness of the proposed method, which mitigates the dilemma of the energy--accuracy trade-off.
comment: 19 pages, 5 figures
☆ Gradient Estimation for Unseen Domain Risk Minimization with Pre-Trained Models
Domain generalization aims to build generalized models that perform well on unseen domains when only source domains are available for model optimization. Recent studies have demonstrated that large-scale pre-trained models could play an important role in domain generalization by providing their generalization power. However, large-scale pre-trained models are not fully equipped with target task-specific knowledge due to a discrepancy between the pre-training objective and the target task. Although the task-specific knowledge could be learned from source domains by fine-tuning, this hurts the generalization power of the pre-trained models because of gradient bias toward the source domains. To address this issue, we propose a new domain generalization method that estimates unobservable gradients that reduce potential risks in unseen domains, using a large-scale pre-trained model. Our proposed method allows the pre-trained model to learn task-specific knowledge further while preserving its generalization ability with the estimated gradients. Experimental results show that our proposed method outperforms baseline methods on DomainBed, a standard benchmark in domain generalization. We also provide extensive analyses to demonstrate that the estimated unobserved gradients relieve the gradient bias, and the pre-trained model learns the task-specific knowledge without sacrificing its generalization power.
☆ Efficient Domain Adaptation for Speech Foundation Models
Foundation models (FMs), that are trained on broad data at scale and are adaptable to a wide range of downstream tasks, have brought large interest in the research community. Benefiting from the diverse data sources such as different modalities, languages and application domains, foundation models have demonstrated strong generalization and knowledge transfer capabilities. In this paper, we present a pioneering study towards building an efficient solution for FM-based speech recognition systems. We adopt the recently developed self-supervised BEST-RQ for pretraining, and propose the joint finetuning with both source and unsupervised target domain data using JUST Hydra. The FM encoder adapter and decoder are then finetuned to the target domain with a small amount of supervised in-domain data. On a large-scale YouTube and Voice Search task, our method is shown to be both data and model parameter efficient. It achieves the same quality with only 21.6M supervised in-domain data and 130.8M finetuned parameters, compared to the 731.1M model trained from scratch on additional 300M supervised in-domain data.
☆ Perfect Is the Enemy of Test Oracle
Automation of test oracles is one of the most challenging facets of software testing, but remains comparatively less addressed compared to automated test input generation. Test oracles rely on a ground-truth that can distinguish between the correct and buggy behavior to determine whether a test fails (detects a bug) or passes. What makes the oracle problem challenging and undecidable is the assumption that the ground-truth should know the exact expected, correct, or buggy behavior. However, we argue that one can still build an accurate oracle without knowing the exact correct or buggy behavior, but how these two might differ. This paper presents SEER, a learning-based approach that in the absence of test assertions or other types of oracle, can determine whether a unit test passes or fails on a given method under test (MUT). To build the ground-truth, SEER jointly embeds unit tests and the implementation of MUTs into a unified vector space, in such a way that the neural representation of tests are similar to that of MUTs they pass on them, but dissimilar to MUTs they fail on them. The classifier built on top of this vector representation serves as the oracle to generate "fail" labels, when test inputs detect a bug in MUT or "pass" labels, otherwise. Our extensive experiments on applying SEER to more than 5K unit tests from a diverse set of open-source Java projects show that the produced oracle is (1) effective in predicting the fail or pass labels, achieving an overall accuracy, precision, recall, and F1 measure of 93%, 86%, 94%, and 90%, (2) generalizable, predicting the labels for the unit test of projects that were not in training or validation set with negligible performance drop, and (3) efficient, detecting the existence of bugs in only 6.5 milliseconds on average.
comment: Published in ESEC/FSE 2022
☆ Xtal2DoS: Attention-based Crystal to Sequence Learning for Density of States Prediction NeurIPS 2022
Modern machine learning techniques have been extensively applied to materials science, especially for property prediction tasks. A majority of these methods address scalar property predictions, while more challenging spectral properties remain less emphasized. We formulate a crystal-to-sequence learning task and propose a novel attention-based learning method, Xtal2DoS, which decodes the sequential representation of the material density of states (DoS) properties by incorporating the learned atomic embeddings through attention networks. Experiments show Xtal2DoS is faster than the existing models, and consistently outperforms other state-of-the-art methods on four metrics for two fundamental spectral properties, phonon and electronic DoS.
comment: Accepted to NeurIPS 2022 AI for Science Workshop
☆ SPADE: Self-supervised Pretraining for Acoustic DisEntanglement
Self-supervised representation learning approaches have grown in popularity due to the ability to train models on large amounts of unlabeled data and have demonstrated success in diverse fields such as natural language processing, computer vision, and speech. Previous self-supervised work in the speech domain has disentangled multiple attributes of speech such as linguistic content, speaker identity, and rhythm. In this work, we introduce a self-supervised approach to disentangle room acoustics from speech and use the acoustic representation on the downstream task of device arbitration. Our results demonstrate that our proposed approach significantly improves performance over a baseline when labeled training data is scarce, indicating that our pretraining scheme learns to encode room acoustic information while remaining invariant to other attributes of the speech signal.
☆ Clustered Embedding Learning for Recommender Systems
In recent years, recommender systems have advanced rapidly, where embedding learning for users and items plays a critical role. A standard method learns a unique embedding vector for each user and item. However, such a method has two important limitations in real-world applications: 1) it is hard to learn embeddings that generalize well for users and items with rare interactions on their own; and 2) it may incur unbearably high memory costs when the number of users and items scales up. Existing approaches either can only address one of the limitations or have flawed overall performances. In this paper, we propose Clustered Embedding Learning (CEL) as an integrated solution to these two problems. CEL is a plug-and-play embedding learning framework that can be combined with any differentiable feature interaction model. It is capable of achieving improved performance, especially for cold users and items, with reduced memory cost. CEL enables automatic and dynamic clustering of users and items in a top-down fashion, where clustered entities jointly learn a shared embedding. The accelerated version of CEL has an optimal time complexity, which supports efficient online updates. Theoretically, we prove the identifiability and the existence of a unique optimal number of clusters for CEL in the context of nonnegative matrix factorization. Empirically, we validate the effectiveness of CEL on three public datasets and one business dataset, showing its consistently superior performance against current state-of-the-art methods. In particular, when incorporating CEL into the business model, it brings an improvement of $+0.6\%$ in AUC, which translates into a significant revenue gain; meanwhile, the size of the embedding table gets $2650$ times smaller.
☆ A Reduction-based Framework for Sequential Decision Making with Delayed Feedback
We study stochastic delayed feedback in general multi-agent sequential decision making, which includes bandits, single-agent Markov decision processes (MDPs), and Markov games (MGs). We propose a novel reduction-based framework, which turns any multi-batched algorithm for sequential decision making with instantaneous feedback into a sample-efficient algorithm that can handle stochastic delays in sequential decision making. By plugging different multi-batched algorithms into our framework, we provide several examples demonstrating that our framework not only matches or improves existing results for bandits, tabular MDPs, and tabular MGs, but also provides the first line of studies on delays in sequential decision making with function approximation. In summary, we provide a complete set of sharp results for multi-agent sequential decision making with delayed feedback.
comment: arXiv admin note: text overlap with arXiv:2110.14555 by other authors
☆ Defensive ML: Defending Architectural Side-channels with Adversarial Obfuscation ICML 2023
Side-channel attacks that use machine learning (ML) for signal analysis have become prominent threats to computer security, as ML models easily find patterns in signals. To address this problem, this paper explores using Adversarial Machine Learning (AML) methods as a defense at the computer architecture layer to obfuscate side channels. We call this approach Defensive ML, and the generator to obfuscate signals, defender. Defensive ML is a workflow to design, implement, train, and deploy defenders for different environments. First, we design a defender architecture given the physical characteristics and hardware constraints of the side-channel. Next, we use our DefenderGAN structure to train the defender. Finally, we apply defensive ML to thwart two side-channel attacks: one based on memory contention and the other on application power. The former uses a hardware defender with ns-level response time that attains a high level of security with half the performance impact of a traditional scheme; the latter uses a software defender with ms-level response time that provides better security than a traditional scheme with only 70% of its power overhead.
comment: Submitted to ICML 2023
☆ User-centric Heterogeneous-action Deep Reinforcement Learning for Virtual Reality in the Metaverse over Wireless Networks
The Metaverse is emerging as maturing technologies are empowering the different facets. Virtual Reality (VR) technologies serve as the backbone of the virtual universe within the Metaverse to offer a highly immersive user experience. As mobility is emphasized in the Metaverse context, VR devices reduce their weights at the sacrifice of local computation abilities. In this paper, for a system consisting of a Metaverse server and multiple VR users, we consider two cases of (i) the server generating frames and transmitting them to users, and (ii) users generating frames locally and thus consuming device energy. Moreover, in our multi-user VR scenario for the Metaverse, users have different characteristics and demands for Frames Per Second (FPS). Then the channel access arrangement (including the decisions on frame generation location), and transmission powers for the downlink communications from the server to the users are jointly optimized to improve the utilities of users. This joint optimization is addressed by deep reinforcement learning (DRL) with heterogeneous actions. Our proposed user-centric DRL algorithm is called User-centric Critic with Heterogenous Actors (UCHA). Extensive experiments demonstrate that our UCHA algorithm leads to remarkable results under various requirements and constraints.
comment: The paper has been submitted to IEEE Transactions on Wireless Communications, 2023
☆ Learning to Optimize for Reinforcement Learning
In recent years, by leveraging more data, computation, and diverse tasks, learned optimizers have achieved remarkable success in supervised learning optimization, outperforming classical hand-designed optimizers. However, in practice, these learned optimizers fail to generalize to reinforcement learning tasks due to unstable and complex loss landscapes. Moreover, neither hand-designed optimizers nor learned optimizers have been specifically designed to address the unique optimization properties in reinforcement learning. In this work, we take a data-driven approach to learn to optimize for reinforcement learning using meta-learning. We introduce a novel optimizer structure that significantly improves the training efficiency of learned optimizers, making it possible to learn an optimizer for reinforcement learning from scratch. Although trained in toy tasks, our learned optimizer demonstrates its generalization ability to unseen complex tasks. Finally, we design a set of small gridworlds to train the first general-purpose optimizer for reinforcement learning.
♻ ☆ Vicarious Offense and Noise Audit of Offensive Speech Classifiers
This paper examines social web content moderation from two key perspectives: automated methods (machine moderators) and human evaluators (human moderators). We conduct a noise audit at an unprecedented scale using nine machine moderators trained on well-known offensive speech data sets evaluated on a corpus sampled from 92 million YouTube comments discussing a multitude of issues relevant to US politics. We introduce a first-of-its-kind data set of vicarious offense. We ask annotators: (1) if they find a given social media post offensive; and (2) how offensive annotators sharing different political beliefs would find the same content. Our experiments with machine moderators reveal that moderation outcomes wildly vary across different machine moderators. Our experiments with human moderators suggest that (1) political leanings considerably affect first-person offense perspective; (2) Republicans are the worst predictors of vicarious offense; (3) predicting vicarious offense for the Republicans is most challenging than predicting vicarious offense for the Independents and the Democrats; and (4) disagreement across political identity groups considerably increases when sensitive issues such as reproductive rights or gun control/rights are discussed. Both experiments suggest that offense, is indeed, highly subjective and raise important questions concerning content moderation practices.
♻ ☆ New Machine Learning Techniques for Simulation-Based Inference: InferoStatic Nets, Kernel Score Estimation, and Kernel Likelihood Ratio Estimation
We propose an intuitive, machine-learning approach to multiparameter inference, dubbed the InferoStatic Networks (ISN) method, to model the score and likelihood ratio estimators in cases when the probability density can be sampled but not computed directly. The ISN uses a backend neural network that models a scalar function called the inferostatic potential $\varphi$. In addition, we introduce new strategies, respectively called Kernel Score Estimation (KSE) and Kernel Likelihood Ratio Estimation (KLRE), to learn the score and the likelihood ratio functions from simulated data. We illustrate the new techniques with some toy examples and compare to existing approaches in the literature. We mention en passant some new loss functions that optimally incorporate latent information from simulations into the training procedure.
comment: 36 pages, 10 figures. Submission to SciPost
♻ ☆ Transformers over Directed Acyclic Graphs
Transformer models have recently gained popularity in graph representation learning as they have the potential to learn complex relationships beyond the ones captured by regular graph neural networks. The main research question is how to inject the structural bias of graphs into the transformer architecture, and several proposals have been made for undirected molecular graphs and, recently, also for larger network graphs. In this paper, we study transformers over directed acyclic graphs (DAGs) and propose architecture adaptations tailored to DAGs: (1) An attention mechanism that is more efficient than the regular quadratic complexity of transformers and at the same time faithfully captures the DAG structure, and (2) a positional encoding of the DAG's partial order, complementing the former. We rigorously evaluate our framework in ablation studies and show that it is effective in improving different kinds of baseline transformers over various types of data, in experiments ranging from classifying source code graphs to nodes in self-citation networks. In particular, our proposal makes (graph) transformers competitive to or outperform graph neural networks tailored to DAGs.
♻ ☆ One-shot domain adaptation in video-based assessment of surgical skills
Deep Learning (DL) has achieved automatic and objective assessment of surgical skills. However, DL models are data-hungry and restricted to their training domain. This prevents them from transitioning to new tasks where data is limited. Hence, domain adaptation is crucial to implement DL in real life. Here, we propose a meta-learning model, A-VBANet, that can deliver domain-agnostic surgical skill classification via one-shot learning. We develop the A-VBANet on five laparoscopic and robotic surgical simulators. Additionally, we test it on operating room (OR) videos of laparoscopic cholecystectomy. Our model successfully adapts with accuracies up to 99.5% in one-shot and 99.9% in few-shot settings for simulated tasks and 89.7% for laparoscopic cholecystectomy. For the first time, we provide a domain-agnostic procedure for video-based assessment of surgical skills. A significant implication of this approach is that it allows the use of data from surgical simulators to assess performance in the operating room.
comment: 12 pages (+9 pages of Supplementary Materials), 4 figures (+2 Supplementary Figures), 2 tables (+5 Supplementary Tables)
♻ ☆ SemEval 2023 Task 9: Multilingual Tweet Intimacy Analysis SemEval 2023
We propose MINT, a new Multilingual INTimacy analysis dataset covering 13,372 tweets in 10 languages including English, French, Spanish, Italian, Portuguese, Korean, Dutch, Chinese, Hindi, and Arabic. We benchmarked a list of popular multilingual pre-trained language models. The dataset is released along with the SemEval 2023 Task 9: Multilingual Tweet Intimacy Analysis (https://sites.google.com/umich.edu/semeval-2023-tweet-intimacy).
comment: SemEval 2023 Task 9: Multilingual Tweet Intimacy Analysis
♻ ☆ MetaNO: How to Transfer Your Knowledge on Learning Hidden Physics
Gradient-based meta-learning methods have primarily been applied to classical machine learning tasks such as image classification. Recently, PDE-solving deep learning methods, such as neural operators, are starting to make an important impact on learning and predicting the response of a complex physical system directly from observational data. Since the data acquisition in this context is commonly challenging and costly, the call of utilization and transfer of existing knowledge to new and unseen physical systems is even more acute. Herein, we propose a novel meta-learning approach for neural operators, which can be seen as transferring the knowledge of solution operators between governing (unknown) PDEs with varying parameter fields. Our approach is a provably universal solution operator for multiple PDE solving tasks, with a key theoretical observation that underlying parameter fields can be captured in the first layer of neural operator models, in contrast to typical final-layer transfer in existing meta-learning methods. As applications, we demonstrate the efficacy of our proposed approach on PDE-based datasets and a real-world material modeling problem, illustrating that our method can handle complex and nonlinear physical response learning tasks while greatly improving the sampling efficiency in unseen tasks.
♻ ☆ Software Simulation and Visualization of Quantum Multi-Drone Reinforcement Learning
Quantum machine learning (QML) has received a lot of attention according to its light training parameter numbers and speeds; and the advances of QML lead to active research on quantum multi-agent reinforcement learning (QMARL). Existing classical multi-agent reinforcement learning (MARL) features non-stationarity and uncertain properties. Therefore, this paper presents a simulation software framework for novel QMARL to control autonomous multi-drones, i.e., quantum multi-drone reinforcement learning. Our proposed framework accomplishes reasonable reward convergence and service quality performance with fewer trainable parameters. Furthermore, it shows more stable training results. Lastly, our proposed software allows us to analyze the training process and results.
comment: Revise paper
♻ ☆ Learning finite difference methods for reaction-diffusion type equations with FCNN
In recent years, Physics-informed neural networks (PINNs) have been widely used to solve partial differential equations alongside numerical methods because PINNs can be trained without observations and deal with continuous-time problems directly. In contrast, optimizing the parameters of such models is difficult, and individual training sessions must be performed to predict the evolutions of each different initial condition. To alleviate the first problem, observed data can be injected directly into the loss function part. To solve the second problem, a network architecture can be built as a framework to learn a finite difference method. In view of the two motivations, we propose Five-point stencil CNNs (FCNNs) containing a five-point stencil kernel and a trainable approximation function for reaction-diffusion type equations including the heat, Fisher's, Allen-Cahn, and other reaction-diffusion equations with trigonometric function terms. We show that FCNNs can learn finite difference schemes using few data and achieve the low relative errors of diverse reaction-diffusion evolutions with unseen initial conditions. Furthermore, we demonstrate that FCNNs can still be trained well even with using noisy data.
comment: 9 figures
♻ ☆ Variational Latent Branching Model for Off-Policy Evaluation ICLR 2023
Model-based methods have recently shown great potential for off-policy evaluation (OPE); offline trajectories induced by behavioral policies are fitted to transitions of Markov decision processes (MDPs), which are used to rollout simulated trajectories and estimate the performance of policies. Model-based OPE methods face two key challenges. First, as offline trajectories are usually fixed, they tend to cover limited state and action space. Second, the performance of model-based methods can be sensitive to the initialization of their parameters. In this work, we propose the variational latent branching model (VLBM) to learn the transition function of MDPs by formulating the environmental dynamics as a compact latent space, from which the next states and rewards are then sampled. Specifically, VLBM leverages and extends the variational inference framework with the recurrent state alignment (RSA), which is designed to capture as much information underlying the limited training data, by smoothing out the information flow between the variational (encoding) and generative (decoding) part of VLBM. Moreover, we also introduce the branching architecture to improve the model's robustness against randomly initialized model weights. The effectiveness of the VLBM is evaluated on the deep OPE (DOPE) benchmark, from which the training trajectories are designed to result in varied coverage of the state-action space. We show that the VLBM outperforms existing state-of-the-art OPE methods in general.
comment: Accepted to ICLR 2023
♻ ☆ Discovering Policies with DOMiNO: Diversity Optimization Maintaining Near Optimality
Finding different solutions to the same problem is a key aspect of intelligence associated with creativity and adaptation to novel situations. In reinforcement learning, a set of diverse policies can be useful for exploration, transfer, hierarchy, and robustness. We propose DOMiNO, a method for Diversity Optimization Maintaining Near Optimality. We formalize the problem as a Constrained Markov Decision Process where the objective is to find diverse policies, measured by the distance between the state occupancies of the policies in the set, while remaining near-optimal with respect to the extrinsic reward. We demonstrate that the method can discover diverse and meaningful behaviors in various domains, such as different locomotion patterns in the DeepMind Control Suite. We perform extensive analysis of our approach, compare it with other multi-objective baselines, demonstrate that we can control both the quality and the diversity of the set via interpretable hyperparameters, and show that the discovered set is robust to perturbations.
♻ ☆ Generating Novel, Designable, and Diverse Protein Structures by Equivariantly Diffusing Oriented Residue Clouds
Proteins power a vast array of functional processes in living cells. The capability to create new proteins with designed structures and functions would thus enable the engineering of cellular behavior and development of protein-based therapeutics and materials. Structure-based protein design aims to find structures that are designable (can be realized by a protein sequence), novel (have dissimilar geometry from natural proteins), and diverse (span a wide range of geometries). While advances in protein structure prediction have made it possible to predict structures of novel protein sequences, the combinatorially large space of sequences and structures limits the practicality of search-based methods. Generative models provide a compelling alternative, by implicitly learning the low-dimensional structure of complex data distributions. Here, we leverage recent advances in denoising diffusion probabilistic models and equivariant neural networks to develop Genie, a generative model of protein structures that performs discrete-time diffusion using a cloud of oriented reference frames in 3D space. Through in silico evaluations, we demonstrate that Genie generates protein backbones that are more designable, novel, and diverse than existing models. This indicates that Genie is capturing key aspects of the distribution of protein structure space and facilitates protein design with high success rates. Code for generating new proteins and training new versions of Genie is available at https://github.com/aqlaboratory/genie.
♻ ☆ On the Tradeoff between Energy, Precision, and Accuracy in Federated Quantized Neural Networks
Deploying federated learning (FL) over wireless networks with resource-constrained devices requires balancing between accuracy, energy efficiency, and precision. Prior art on FL often requires devices to train deep neural networks (DNNs) using a 32-bit precision level for data representation to improve accuracy. However, such algorithms are impractical for resource-constrained devices since DNNs could require execution of millions of operations. Thus, training DNNs with a high precision level incurs a high energy cost for FL. In this paper, a quantized FL framework, that represents data with a finite level of precision in both local training and uplink transmission, is proposed. Here, the finite level of precision is captured through the use of quantized neural networks (QNNs) that quantize weights and activations in fixed-precision format. In the considered FL model, each device trains its QNN and transmits a quantized training result to the base station. Energy models for the local training and the transmission with the quantization are rigorously derived. An energy minimization problem is formulated with respect to the level of precision while ensuring convergence. To solve the problem, we first analytically derive the FL convergence rate and use a line search method. Simulation results show that our FL framework can reduce energy consumption by up to 53% compared to a standard FL model. The results also shed light on the tradeoff between precision, energy, and accuracy in FL over wireless networks.
comment: This paper is accepted by IEEE International Conference on Communications 2022
♻ ☆ MonoFlow: Rethinking Divergence GANs via the Perspective of Differential Equations
The conventional understanding of adversarial training in generative adversarial networks (GANs) is that the discriminator is trained to estimate a divergence, and the generator learns to minimize this divergence. We argue that despite the fact that many variants of GANs were developed following this paradigm, the current theoretical understanding of GANs and their practical algorithms are inconsistent. In this paper, we leverage Wasserstein gradient flows which characterize the evolution of particles in the sample space, to gain theoretical insights and algorithmic inspiration of GANs. We introduce a unified generative modeling framework - MonoFlow: the particle evolution is rescaled via a monotonically increasing mapping of the log density ratio. Under our framework, adversarial training can be viewed as a procedure first obtaining MonoFlow's vector field via training the discriminator and the generator learns to draw the particle flow defined by the corresponding vector field. We also reveal the fundamental difference between variational divergence minimization and adversarial training. This analysis helps us to identify what types of generator loss functions can lead to the successful training of GANs and suggest that GANs may have more loss designs beyond the literature (e.g., non-saturated loss), as long as they realize MonoFlow. Consistent empirical studies are included to validate the effectiveness of our framework.
♻ ☆ Anti-Symmetric DGN: a stable architecture for Deep Graph Networks ICLR 2023
Deep Graph Networks (DGNs) currently dominate the research landscape of learning from graphs, due to their efficiency and ability to implement an adaptive message-passing scheme between the nodes. However, DGNs are typically limited in their ability to propagate and preserve long-term dependencies between nodes, i.e., they suffer from the over-squashing phenomena. This reduces their effectiveness, since predictive problems may require to capture interactions at different, and possibly large, radii in order to be effectively solved. In this work, we present Anti-Symmetric Deep Graph Networks (A-DGNs), a framework for stable and non-dissipative DGN design, conceived through the lens of ordinary differential equations. We give theoretical proof that our method is stable and non-dissipative, leading to two key results: long-range information between nodes is preserved, and no gradient vanishing or explosion occurs in training. We empirically validate the proposed approach on several graph benchmarks, showing that A-DGN yields to improved performance and enables to learn effectively even when dozens of layers are used.
comment: Accepted at ICLR 2023 (https://openreview.net/forum?id=J3Y7cgZOOS)
♻ ☆ Convolutional Autoencoders, Clustering and POD for Low-dimensional Parametrization of Navier-Stokes Equations
Simulations of large-scale dynamical systems require expensive computations. Low-dimensional parametrization of high-dimensional states such as Proper Orthogonal Decomposition (POD) can be a solution to lessen the burdens by providing a certain compromise between accuracy and model complexity. However, for really low-dimensional parametrizations (for example for controller design) linear methods like the POD come to their natural limits so that nonlinear approaches will be the methods of choice. In this work we propose a convolutional autoencoder (CAE) consisting of a nonlinear encoder and an affine linear decoder and consider combinations with k-means clustering for improved encoding performance. The proposed set of methods is compared to the standard POD approach in two cylinder-wake scenarios modeled by the incompressible Navier-Stokes equations.
comment: 16 pages, 12 figures
♻ ☆ Testing Rare Downstream Safety Violations via Upstream Adaptive Sampling of Perception Error Models
Testing black-box perceptual-control systems in simulation faces two difficulties. Firstly, perceptual inputs in simulation lack the fidelity of real-world sensor inputs. Secondly, for a reasonably accurate perception system, encountering a rare failure trajectory may require running infeasibly many simulations. This paper combines perception error models -- surrogates for a sensor-based detection system -- with state-dependent adaptive importance sampling. This allows us to efficiently assess the rare failure probabilities for real-world perceptual control systems within simulation. Our experiments with an autonomous braking system equipped with an RGB obstacle-detector show that our method can calculate accurate failure probabilities with an inexpensive number of simulations. Further, we show how choice of safety metric can influence the process of learning proposal distributions capable of reliably sampling high-probability failures.
♻ ☆ Fast Feature Selection with Fairness Constraints
We study the fundamental problem of selecting optimal features for model construction. This problem is computationally challenging on large datasets, even with the use of greedy algorithm variants. To address this challenge, we extend the adaptive query model, recently proposed for the greedy forward selection for submodular functions, to the faster paradigm of Orthogonal Matching Pursuit for non-submodular functions. The proposed algorithm achieves exponentially fast parallel run time in the adaptive query model, scaling much better than prior work. Furthermore, our extension allows the use of downward-closed constraints, which can be used to encode certain fairness criteria into the feature selection process. We prove strong approximation guarantees for the algorithm based on standard assumptions. These guarantees are applicable to many parametric models, including Generalized Linear Models. Finally, we demonstrate empirically that the proposed algorithm competes favorably with state-of-the-art techniques for feature selection, on real-world and synthetic datasets.
♻ ☆ Learning Counterfactually Invariant Predictors
Counterfactual invariance has proven an essential property for predictors that are fair, robust, and generalizable in the real world. We propose a general definition of counterfactual invariance and provide simple graphical criteria that yield a sufficient condition for a predictor to be counterfactually invariant in terms of (conditional independence in) the observational distribution. Any predictor that satisfies our criterion is provably counterfactually invariant. In order to learn such predictors, we propose a model-agnostic framework, called Counterfactual Invariance Prediction (CIP), based on a kernel-based conditional dependence measure called Hilbert-Schmidt Conditional Independence Criterion (HSCIC). Our experimental results demonstrate the effectiveness of CIP in enforcing counterfactual invariance across various types of data including tabular, high-dimensional, and real-world dataset.
♻ ☆ When Data Geometry Meets Deep Function: Generalizing Offline Reinforcement Learning ICLR2023
In offline reinforcement learning (RL), one detrimental issue to policy learning is the error accumulation of deep Q function in out-of-distribution (OOD) areas. Unfortunately, existing offline RL methods are often over-conservative, inevitably hurting generalization performance outside data distribution. In our study, one interesting observation is that deep Q functions approximate well inside the convex hull of training data. Inspired by this, we propose a new method, DOGE (Distance-sensitive Offline RL with better GEneralization). DOGE marries dataset geometry with deep function approximators in offline RL, and enables exploitation in generalizable OOD areas rather than strictly constraining policy within data distribution. Specifically, DOGE trains a state-conditioned distance function that can be readily plugged into standard actor-critic methods as a policy constraint. Simple yet elegant, our algorithm enjoys better generalization compared to state-of-the-art methods on D4RL benchmarks. Theoretical analysis demonstrates the superiority of our approach to existing methods that are solely based on data distribution or support constraints.
comment: Accepted by ICLR2023
♻ ☆ PDEBENCH: An Extensive Benchmark for Scientific Machine Learning NeurIPS 2022
Machine learning-based modeling of physical systems has experienced increased interest in recent years. Despite some impressive progress, there is still a lack of benchmarks for Scientific ML that are easy to use but still challenging and representative of a wide range of problems. We introduce PDEBench, a benchmark suite of time-dependent simulation tasks based on Partial Differential Equations (PDEs). PDEBench comprises both code and data to benchmark the performance of novel machine learning models against both classical numerical simulations and machine learning baselines. Our proposed set of benchmark problems contribute the following unique features: (1) A much wider range of PDEs compared to existing benchmarks, ranging from relatively common examples to more realistic and difficult problems; (2) much larger ready-to-use datasets compared to prior work, comprising multiple simulation runs across a larger number of initial and boundary conditions and PDE parameters; (3) more extensible source codes with user-friendly APIs for data generation and baseline results with popular machine learning models (FNO, U-Net, PINN, Gradient-Based Inverse Method). PDEBench allows researchers to extend the benchmark freely for their own purposes using a standardized API and to compare the performance of new models to existing baseline methods. We also propose new evaluation metrics with the aim to provide a more holistic understanding of learning methods in the context of Scientific ML. With those metrics we identify tasks which are challenging for recent ML methods and propose these tasks as future challenges for the community. The code is available at https://github.com/pdebench/PDEBench.
comment: 16 pages (main body) + 34 pages (supplemental material), accepted for publication in NeurIPS 2022 Track Datasets and Benchmarks
♻ ☆ Causal Modeling of Policy Interventions From Sequences of Treatments and Outcomes using Gaussian Processes
A treatment policy defines when and what treatments are applied to affect some outcome of interest. Data-driven decision-making requires the ability to predict what happens if a policy is changed. Existing methods that predict how the outcome evolves under different scenarios assume that the tentative sequences of future treatments are fixed in advance, while in practice the treatments are determined stochastically by a policy and may depend for example on the efficiency of previous treatments. Therefore, the current methods are not applicable if the treatment policy is unknown or a counterfactual analysis is needed. To handle these limitations, we model the treatments and outcomes jointly in continuous time, by combining Gaussian processes and point processes. Our model enables the estimation of a treatment policy from observational sequences of treatments and outcomes, and it can predict the interventional and counterfactual progression of the outcome after an intervention on the treatment policy (in contrast with the causal effect of a single treatment). We show with real-world and semi-synthetic data on blood glucose progression that our method can answer causal queries more accurately than existing alternatives.
♻ ☆ Realizable Learning is All You Need
The equivalence of realizable and agnostic learnability is a fundamental phenomenon in learning theory. With variants ranging from classical settings like PAC learning and regression to recent trends such as adversarially robust learning, it's surprising that we still lack a unified theory; traditional proofs of the equivalence tend to be disparate, and rely on strong model-specific assumptions like uniform convergence and sample compression. In this work, we give the first model-independent framework explaining the equivalence of realizable and agnostic learnability: a three-line blackbox reduction that simplifies, unifies, and extends our understanding across a wide variety of settings. This includes models with no known characterization of learnability such as learning with arbitrary distributional assumptions and more general loss functions, as well as a host of other popular settings such as robust learning, partial learning, fair learning, and the statistical query model. More generally, we argue that the equivalence of realizable and agnostic learning is actually a special case of a broader phenomenon we call property generalization: any desirable property of a learning algorithm (e.g. noise tolerance, privacy, stability) that can be satisfied over finite hypothesis classes extends (possibly in some variation) to any learnable hypothesis class.
♻ ☆ Socially Fair Reinforcement Learning
We consider the problem of episodic reinforcement learning where there are multiple stakeholders with different reward functions. Our goal is to output a policy that is socially fair with respect to different reward functions. Prior works have proposed different objectives that a fair policy must optimize including minimum welfare, and generalized Gini welfare. We first take an axiomatic view of the problem, and propose four axioms that any such fair objective must satisfy. We show that the Nash social welfare is the unique objective that uniquely satisfies all four objectives, whereas prior objectives fail to satisfy all four axioms. We then consider the learning version of the problem where the underlying model i.e. Markov decision process is unknown. We consider the problem of minimizing regret with respect to the fair policies maximizing three different fair objectives -- minimum welfare, generalized Gini welfare, and Nash social welfare. Based on optimistic planning, we propose a generic learning algorithm and derive its regret bound with respect to the three different policies. For the objective of Nash social welfare, we also derive a lower bound in regret that grows exponentially with $n$, the number of agents. Finally, we show that for the objective of minimum welfare, one can improve regret by a factor of $O(H)$ for a weaker notion of regret.
♻ ☆ Data Representativity for Machine Learning and AI Systems
Data representativity is crucial when drawing inference from data through machine learning models. Scholars have increased focus on unraveling the bias and fairness in models, also in relation to inherent biases in the input data. However, limited work exists on the representativity of samples (datasets) for appropriate inference in AI systems. This paper reviews definitions and notions of a representative sample and surveys their use in scientific AI literature. We introduce three measurable concepts to help focus the notions and evaluate different data samples. Furthermore, we demonstrate that the contrast between a representative sample in the sense of coverage of the input space, versus a representative sample mimicking the distribution of the target population is of particular relevance when building AI systems. Through empirical demonstrations on US Census data, we evaluate the opposing inherent qualities of these concepts. Finally, we propose a framework of questions for creating and documenting data with data representativity in mind, as an addition to existing dataset documentation templates.
♻ ☆ Searching for the Essence of Adversarial Perturbations
Neural networks have demonstrated state-of-the-art performance in various machine learning fields. However, the introduction of malicious perturbations in input data, known as adversarial examples, has been shown to deceive neural network predictions. This poses potential risks for real-world applications such as autonomous driving and text identification. In order to mitigate these risks, a comprehensive understanding of the mechanisms underlying adversarial examples is essential. In this study, we demonstrate that adversarial perturbations contain human-recognizable information, which is the key conspirator responsible for a neural network's incorrect prediction, in contrast to the widely held belief that human-unidentifiable characteristics play a critical role in fooling a network. This concept of human-recognizable characteristics enables us to explain key features of adversarial perturbations, including their existence, transferability among different neural networks, and increased interpretability for adversarial training. We also uncover two unique properties of adversarial perturbations that deceive neural networks: masking and generation. Additionally, a special class, the complementary class, is identified when neural networks classify input images. The presence of human-recognizable information in adversarial perturbations allows researchers to gain insight into the working principles of neural networks and may lead to the development of techniques for detecting and defending against adversarial attacks.
♻ ☆ Graph Neural Networks for temporal graphs: State of the art, open challenges, and opportunities
Graph Neural Networks (GNNs) have become the leading paradigm for learning on (static) graph-structured data. However, many real-world systems are dynamic in nature, since the graph and node/edge attributes change over time. In recent years, GNN-based models for temporal graphs have emerged as a promising area of research to extend the capabilities of GNNs. In this work, we provide the first comprehensive overview of the current state-of-the-art of temporal GNN, introducing a rigorous formalization of learning settings and tasks and a novel taxonomy categorizing existing approaches in terms of how the temporal aspect is represented and processed. We conclude the survey with a discussion of the most relevant open challenges for the field, from both research and application perspectives.
♻ ☆ IR-MCL: Implicit Representation-Based Online Global Localization
Determining the state of a mobile robot is an essential building block of robot navigation systems. In this paper, we address the problem of estimating the robots pose in an indoor environment using 2D LiDAR data and investigate how modern environment models can improve gold standard Monte-Carlo localization (MCL) systems. We propose a neural occupancy field to implicitly represent the scene using a neural network. With the pretrained network, we can synthesize 2D LiDAR scans for an arbitrary robot pose through volume rendering. Based on the implicit representation, we can obtain the similarity between a synthesized and actual scan as an observation model and integrate it into an MCL system to perform accurate localization. We evaluate our approach on self-recorded datasets and three publicly available ones. We show that we can accurately and efficiently localize a robot using our approach surpassing the localization performance of state-of-the-art methods. The experiments suggest that the presented implicit representation is able to predict more accurate 2D LiDAR scans leading to an improved observation model for our particle filter-based localization. The code of our approach will be available at: https://github.com/PRBonn/ir-mcl.
comment: 8 pages, 5 figures. Accepted to IEEE Robotics and Automation Letters
♻ ☆ Physically Consistent Learning of Conservative Lagrangian Systems with Gaussian Processes
This paper proposes a physically consistent Gaussian Process (GP) enabling the identification of uncertain Lagrangian systems. The function space is tailored according to the energy components of the Lagrangian and the differential equation structure, analytically guaranteeing physical and mathematical properties such as energy conservation and quadratic form. The novel formulation of Cholesky decomposed matrix kernels allow the probabilistic preservation of positive definiteness. Only differential input-to-output measurements of the function map are required while Gaussian noise is permitted in torques, velocities, and accelerations. We demonstrate the effectiveness of the approach in numerical simulation.
comment: Accepted version of paper published by IEEE in 2022 IEEE 61st Conference on Decision and Control (CDC). Final published paper can be found at https://doi.org/10.1109/CDC51059.2022.9993123
♻ ☆ Predicting Molecule-Target Interaction by Learning Biomedical Network and Molecule Representations
The study of molecule-target interaction is quite important for drug discovery in terms of target identification, hit identification, pathway study, drug-drug interaction, etc. Most existing methodologies utilize either biomedical network information or molecule structural features to predict potential interaction link. However, the biomedical network information based methods usually suffer from cold start problem, while structure based methods often give limited performance due to the structure/interaction assumption and data quality. To address these issues, we propose a pseudo-siamese Graph Neural Network method, namely MTINet+, which learns both biomedical network topological and molecule structural/chemical information as representations to predict potential interaction of given molecule and target pair. In MTINet+, 1-hop subgraphs of given molecule and target pair are extracted from known interaction of biomedical network as topological information, meanwhile the molecule structural and chemical attributes are processed as molecule information. MTINet+ learns these two types of information as embedding features for predicting the pair link. In the experiments of different molecule-target interaction tasks, MTINet+ significantly outperforms over the state-of-the-art baselines. In addition, in our designed network sparsity experiments , MTINet+ shows strong robustness against different sparse biomedical networks.
comment: 9 pages, 6 figures. arXiv admin note: substantial text overlap with arXiv:2102.01649
♻ ☆ Post-Selection Confidence Bounds for Prediction Performance
In machine learning, the selection of a promising model from a potentially large number of competing models and the assessment of its generalization performance are critical tasks that need careful consideration. Typically, model selection and evaluation are strictly separated endeavors, splitting the sample at hand into a training, validation, and evaluation set, and only compute a single confidence interval for the prediction performance of the final selected model. We however propose an algorithm how to compute valid lower confidence bounds for multiple models that have been selected based on their prediction performances in the evaluation set by interpreting the selection problem as a simultaneous inference problem. We use bootstrap tilting and a maxT-type multiplicity correction. The approach is universally applicable for any combination of prediction models, any model selection strategy, and any prediction performance measure that accepts weights. We conducted various simulation experiments which show that our proposed approach yields lower confidence bounds that are at least comparably good as bounds from standard approaches, and that reliably reach the nominal coverage probability. In addition, especially when sample size is small, our proposed approach yields better performing prediction models than the default selection of only one model for evaluation does.
comment: 17 pages, 13 figures, 3 tables. Submitted to the Springer Machine Learning Journal. Changes to version 2: made figures easier to read; corrected a minor typo
♻ ☆ Accelerometry-based classification of circulatory states during out-of-hospital cardiac arrest
Objective: Exploit accelerometry data for an automatic, reliable, and prompt detection of spontaneous circulation during cardiac arrest, as this is both vital for patient survival and practically challenging. Methods: We developed a machine learning algorithm to automatically predict the circulatory state during cardiopulmonary resuscitation from 4-second-long snippets of accelerometry and electrocardiogram (ECG) data from pauses of chest compressions of real-world defibrillator records. The algorithm was trained based on 422 cases from the German Resuscitation Registry, for which ground truth labels were created by a manual annotation of physicians. It uses a kernelized Support Vector Machine classifier based on 49 features, which partially reflect the correlation between accelerometry and electrocardiogram data. Results: Evaluating 50 different test-training data splits, the proposed algorithm exhibits a balanced accuracy of 81.2%, a sensitivity of 80.6%, and a specificity of 81.8%, whereas using only ECG leads to a balanced accuracy of 76.5%, a sensitivity of 80.2%, and a specificity of 72.8%. Conclusion: The first method employing accelerometry for pulse/no-pulse decision yields a significant increase in performance compared to single ECG-signal usage. Significance: This shows that accelerometry provides relevant information for pulse/no-pulse decisions. In application, such an algorithm may be used to simplify retrospective annotation for quality management and, moreover, to support clinicians to assess circulatory state during cardiac arrest treatment.
comment: 12 pages, 7 figures
♻ ☆ GOOD: Exploring Geometric Cues for Detecting Objects in an Open World ICLR 2023
We address the task of open-world class-agnostic object detection, i.e., detecting every object in an image by learning from a limited number of base object classes. State-of-the-art RGB-based models suffer from overfitting the training classes and often fail at detecting novel-looking objects. This is because RGB-based models primarily rely on appearance similarity to detect novel objects and are also prone to overfitting short-cut cues such as textures and discriminative parts. To address these shortcomings of RGB-based object detectors, we propose incorporating geometric cues such as depth and normals, predicted by general-purpose monocular estimators. Specifically, we use the geometric cues to train an object proposal network for pseudo-labeling unannotated novel objects in the training set. Our resulting Geometry-guided Open-world Object Detector (GOOD) significantly improves detection recall for novel object categories and already performs well with only a few training classes. Using a single "person" class for training on the COCO dataset, GOOD surpasses SOTA methods by 5.0% AR@100, a relative improvement of 24%.
comment: Published as a conference paper at ICLR 2023
♻ ☆ A Systematic Survey of Molecular Pre-trained Models
Deep learning has achieved remarkable success in learning representations for molecules, which is crucial for various biochemical applications, ranging from property prediction to drug design. However, training Deep Neural Networks (DNNs) from scratch often requires abundant labeled molecules, which are expensive to acquire in the real world. To alleviate this issue, tremendous efforts have been devoted to Molecular Pre-trained Models (MPMs), where DNNs are pre-trained using large-scale unlabeled molecular databases and then fine-tuned over specific downstream tasks. Despite the prosperity, there lacks a systematic review of this fast-growing field. In this paper, we present the first survey that summarizes the current progress of MPMs. We first highlight the limitations of training molecular representation models from scratch to motivate MPM studies. Next, we systematically review recent advances on this topic from several key perspectives, including molecular descriptors, encoder architectures, pre-training strategies, and applications. We also highlight the challenges and promising avenues for future research, providing a useful resource for both machine learning and scientific communities.
comment: 9 pages, work in progress
♻ ☆ A*Net: A Scalable Path-based Reasoning Approach for Knowledge Graphs
Reasoning on large-scale knowledge graphs has been long dominated by embedding methods. While path-based methods possess the inductive capacity that embeddings lack, they suffer from the scalability issue due to the exponential number of paths. Here we present A*Net, a scalable path-based method for knowledge graph reasoning. Inspired by the A* algorithm for shortest path problems, our A*Net learns a priority function to select important nodes and edges at each iteration, to reduce time and memory footprint for both training and inference. The ratio of selected nodes and edges can be specified to trade off between performance and efficiency. Experiments on both transductive and inductive knowledge graph reasoning benchmarks show that A*Net achieves competitive performance with existing state-of-the-art path-based methods, while merely visiting 10% nodes and 10% edges at each iteration. On a million-scale dataset ogbl-wikikg2, A*Net achieves competitive performance with embedding methods and converges faster. To our best knowledge, A*Net is the first path-based method for knowledge graph reasoning at such a scale.
♻ ☆ Consistent Range Approximation for Fair Predictive Modeling
This paper proposes a novel framework for certifying the fairness of predictive models trained on biased data. It draws from query answering for incomplete and inconsistent databases to formulate the problem of consistent range approximation (CRA) of fairness queries for a predictive model on a target population. The framework employs background knowledge of the data collection process and biased data, working with or without limited statistics about the target population, to compute a range of answers for fairness queries. Using CRA, the framework builds predictive models that are certifiably fair on the target population, regardless of the availability of external data during training. The framework's efficacy is demonstrated through evaluations on real data, showing substantial improvement over existing state-of-the-art methods.
♻ ☆ RL4ReAl: Reinforcement Learning for Register Allocation
We aim to automate decades of research and experience in register allocation, leveraging machine learning. We tackle this problem by embedding a multi-agent reinforcement learning algorithm within LLVM, training it with the state of the art techniques. We formalize the constraints that precisely define the problem for a given instruction-set architecture, while ensuring that the generated code preserves semantic correctness. We also develop a gRPC based framework providing a modular and efficient compiler interface for training and inference. Our approach is architecture independent: we show experimental results targeting Intel x86 and ARM AArch64. Our results match or out-perform the heavily tuned, production-grade register allocators of LLVM.
comment: Published in CC'23
♻ ☆ SimMTM: A Simple Pre-Training Framework for Masked Time-Series Modeling
Time series analysis is widely used in extensive areas. Recently, to reduce labeling expenses and benefit various tasks, self-supervised pre-training has attracted immense interest. One mainstream paradigm is masked modeling, which successfully pre-trains deep models by learning to reconstruct the masked content based on the unmasked part. However, since the semantic information of time series is mainly contained in temporal variations, the standard way of randomly masking a portion of time points will ruin vital temporal variations of time series seriously, making the reconstruction task too difficult to guide representation learning. We thus present SimMTM, a Simple pre-training framework for Masked Time-series Modeling. By relating masked modeling to manifold learning, SimMTM proposes to recover masked time points by the weighted aggregation of multiple neighbors outside the manifold, which eases the reconstruction task by assembling ruined but complementary temporal variations from multiple masked series. SimMTM further learns to uncover the local structure of the manifold helpful for masked modeling. Experimentally, SimMTM achieves state-of-the-art fine-tuning performance in two canonical time series analysis tasks: forecasting and classification, covering both in- and cross-domain settings.
♻ ☆ Language Quantized AutoEncoders: Towards Unsupervised Text-Image Alignment
Recent progress in scaling up large language models has shown impressive capabilities in performing few-shot learning across a wide range of text-based tasks. However, a key limitation is that these language models fundamentally lack visual perception - a crucial attribute needed to extend these models to be able to interact with the real world and solve vision tasks, such as in visual-question answering and robotics. Prior works have largely connected image to text through pretraining and/or fine-tuning on curated image-text datasets, which can be a costly and expensive process. In order to resolve this limitation, we propose a simple yet effective approach called Language-Quantized AutoEncoder (LQAE), a modification of VQ-VAE that learns to align text-image data in an unsupervised manner by leveraging pretrained language models (e.g., BERT, RoBERTa). Our main idea is to encode image as sequences of text tokens by directly quantizing image embeddings using a pretrained language codebook. We then apply random masking followed by a BERT model, and have the decoder reconstruct the original image from BERT predicted text token embeddings. By doing so, LQAE learns to represent similar images with similar clusters of text tokens, thereby aligning these two modalities without the use of aligned text-image pairs. This enables few-shot image classification with large language models (e.g., GPT-3) as well as linear classification of images based on BERT text features. To the best of our knowledge, our work is the first work that uses unaligned images for multimodal tasks by leveraging the power of pretrained language models.
comment: Fixed typos
DEUP: Direct Epistemic Uncertainty Prediction
Epistemic Uncertainty is a measure of the lack of knowledge of a learner which diminishes with more evidence. While existing work focuses on using the variance of the Bayesian posterior due to parameter uncertainty as a measure of epistemic uncertainty, we argue that this does not capture the part of lack of knowledge induced by model misspecification. We discuss how the excess risk, which is the gap between the generalization error of a predictor and the Bayes predictor, is a sound measure of epistemic uncertainty which captures the effect of model misspecification. We thus propose a principled framework for directly estimating the excess risk by learning a secondary predictor for the generalization error and subtracting an estimate of aleatoric uncertainty, i.e., intrinsic unpredictability. We discuss the merits of this novel measure of epistemic uncertainty, and highlight how it differs from variance-based measures of epistemic uncertainty and addresses its major pitfall. Our framework, Direct Epistemic Uncertainty Prediction (DEUP) is particularly interesting in interactive learning environments, where the learner is allowed to acquire novel examples in each round. Through a wide set of experiments, we illustrate how existing methods in sequential model optimization can be improved with epistemic uncertainty estimates from DEUP, and how DEUP can be used to drive exploration in reinforcement learning. We also evaluate the quality of uncertainty estimates from DEUP for probabilistic image classification and predicting synergies of drug combinations.
♻ ☆ Understanding Gradient Regularization in Deep Learning: Efficient Finite-Difference Computation and Implicit Bias
Gradient regularization (GR) is a method that penalizes the gradient norm of the training loss during training. While some studies have reported that GR can improve generalization performance, little attention has been paid to it from the algorithmic perspective, that is, the algorithms of GR that efficiently improve the performance. In this study, we first reveal that a specific finite-difference computation, composed of both gradient ascent and descent steps, reduces the computational cost of GR. Next, we show that the finite-difference computation also works better in the sense of generalization performance. We theoretically analyze a solvable model, a diagonal linear network, and clarify that GR has a desirable implicit bias to so-called rich regime and finite-difference computation strengthens this bias. Furthermore, finite-difference GR is closely related to some other algorithms based on iterative ascent and descent steps for exploring flat minima. In particular, we reveal that the flooding method can perform finite-difference GR in an implicit way. Thus, this work broadens our understanding of GR for both practice and theory.
♻ ☆ On the Opportunity of Causal Deep Generative Models: A Survey and Future Directions
Deep generative models have gained popularity in recent years due to their ability to accurately replicate inherent empirical distributions and yield novel samples. In particular, certain advances are proposed wherein the model engenders data examples following specified attributes. Nevertheless, several challenges still exist and are to be overcome, i.e., difficulty in extrapolating out-of-sample data and insufficient learning of disentangled representations. Structural causal models (SCMs), on the other hand, encapsulate the causal factors that govern a generative process and characterize a generative model based on causal relationships, providing crucial insights for addressing the current obstacles in deep generative models. In this paper, we present a comprehensive survey of Causal deep Generative Models (CGMs), which combine SCMs and deep generative models in a way that boosts several trustworthy properties such as robustness, fairness, and interpretability. We provide an overview of the recent advances in CGMs, categorize them based on generative types, and discuss how causality is introduced into the family of deep generative models. We also explore potential avenues for future research in this field.
♻ ☆ Conditional Antibody Design as 3D Equivariant Graph Translation ICLR 2023
Antibody design is valuable for therapeutic usage and biological research. Existing deep-learning-based methods encounter several key issues: 1) incomplete context for Complementarity-Determining Regions (CDRs) generation; 2) incapability of capturing the entire 3D geometry of the input structure; 3) inefficient prediction of the CDR sequences in an autoregressive manner. In this paper, we propose Multi-channel Equivariant Attention Network (MEAN) to co-design 1D sequences and 3D structures of CDRs. To be specific, MEAN formulates antibody design as a conditional graph translation problem by importing extra components including the target antigen and the light chain of the antibody. Then, MEAN resorts to E(3)-equivariant message passing along with a proposed attention mechanism to better capture the geometrical correlation between different components. Finally, it outputs both the 1D sequences and 3D structure via a multi-round progressive full-shot scheme, which enjoys more efficiency and precision against previous autoregressive approaches. Our method significantly surpasses state-of-the-art models in sequence and structure modeling, antigen-binding CDR design, and binding affinity optimization. Specifically, the relative improvement to baselines is about 23% in antigen-binding CDR design and 34% for affinity optimization.
comment: Accepted to ICLR 2023 as oral presentation
♻ ☆ Fast Bayesian Optimization of Needle-in-a-Haystack Problems using Zooming Memory-Based Initialization (ZoMBI)
Needle-in-a-Haystack problems exist across a wide range of applications including rare disease prediction, ecological resource management, fraud detection, and material property optimization. A Needle-in-a-Haystack problem arises when there is an extreme imbalance of optimum conditions relative to the size of the dataset. For example, only $0.82\%$ out of $146$k total materials in the open-access Materials Project database have a negative Poisson's ratio. However, current state-of-the-art optimization algorithms are not designed with the capabilities to find solutions to these challenging multidimensional Needle-in-a-Haystack problems, resulting in slow convergence to a global optimum or pigeonholing into a local minimum. In this paper, we present a Zooming Memory-Based Initialization algorithm, entitled ZoMBI. ZoMBI actively extracts knowledge from the previously best-performing evaluated experiments to iteratively zoom in the sampling search bounds towards the global optimum "needle" and then prunes the memory of low-performing historical experiments to accelerate compute times by reducing the algorithm time complexity from $O(n^3)$ to $O(\phi^3)$ for $\phi$ forward experiments per activation, which trends to a constant $O(1)$ over several activations. Additionally, ZoMBI implements two custom adaptive acquisition functions to further guide the sampling of new experiments toward the global optimum. We validate the algorithm's optimization performance on three real-world datasets exhibiting Needle-in-a-Haystack and further stress-test the algorithm's performance on an additional 174 analytical datasets. The ZoMBI algorithm demonstrates compute time speed-ups of 400x compared to traditional Bayesian optimization as well as efficiently discovering optima in under 100 experiments that are up to 3x more highly optimized than those discovered by similar methods MiP-EGO, TuRBO, and HEBO.
comment: Paper 16 pages; SI 6 pages
♻ ☆ S$^3$NN: Time Step Reduction of Spiking Surrogate Gradients for Training Energy Efficient Single-Step Spiking Neural Networks
As the scales of neural networks increase, techniques that enable them to run with low computational cost and energy efficiency are required. From such demands, various efficient neural network paradigms, such as spiking neural networks (SNNs) or binary neural networks (BNNs), have been proposed. However, they have sticky drawbacks, such as degraded inference accuracy and latency. To solve these problems, we propose a single-step spiking neural network (S$^3$NN), an energy-efficient neural network with low computational cost and high precision. The proposed S$^3$NN processes the information between hidden layers by spikes as SNNs. Nevertheless, it has no temporal dimension so that there is no latency within training and inference phases as BNNs. Thus, the proposed S$^3$NN has a lower computational cost than SNNs that require time-series processing. However, S$^3$NN cannot adopt na\"{i}ve backpropagation algorithms due to the non-differentiability nature of spikes. We deduce a suitable neuron model by reducing the surrogate gradient for multi-time step SNNs to a single-time step. We experimentally demonstrated that the obtained surrogate gradient allows S$^3$NN to be trained appropriately. We also showed that the proposed S$^3$NN could achieve comparable accuracy to full-precision networks while being highly energy-efficient.
comment: 23 pages, 6 figures
♻ ☆ Lung airway geometry as an early predictor of autism: A preliminary machine learning-based study
The goal of this study is to assess the feasibility of airway geometry as a biomarker for ASD. Chest CT images of children with a documented diagnosis of ASD as well as healthy controls were identified retrospectively. 54 scans were obtained for analysis, including 31 ASD cases and 23 age and sex-matched controls. A feature selection and classification procedure using principal component analysis (PCA) and support vector machine (SVM) achieved a peak cross validation accuracy of nearly 89% using a feature set of 8 airway branching angles. Sensitivity was 94%, but specificity was only 78%. The results suggest a measurable difference in airway branchpoint angles between children with ASD and the control population. Under review at Scientific Reports
♻ ☆ Incorporating Prior Knowledge into Neural Networks through an Implicit Composite Kernel
It is challenging to guide neural network (NN) learning with prior knowledge. In contrast, many known properties, such as spatial smoothness or seasonality, are straightforward to model by choosing an appropriate kernel in a Gaussian process (GP). Many deep learning applications could be enhanced by modeling such known properties. For example, convolutional neural networks (CNNs) are frequently used in remote sensing, which is subject to strong seasonal effects. We propose to blend the strengths of deep learning and the clear modeling capabilities of GPs by using a composite kernel that combines a kernel implicitly defined by a neural network with a second kernel function chosen to model known properties (e.g., seasonality). We implement this idea by combining a deep network and an efficient mapping based on the Nystrom approximation, which we call Implicit Composite Kernel (ICK). We then adopt a sample-then-optimize approach to approximate the full GP posterior distribution. We demonstrate that ICK has superior performance and flexibility on both synthetic and real-world data sets. We believe that ICK framework can be used to include prior information into neural networks in many applications.
comment: 23 pages, 14 figures, 3 tables, 2 algorithms
♻ ☆ Estimating Causal Effects using a Multi-task Deep Ensemble
Over the past few decades, a number of methods have been proposed for causal effect estimation, yet few have been demonstrated to be effective in handling data with complex structures, such as images. To fill this gap, we propose a Causal Multi-task Deep Ensemble (CMDE) framework to learn both shared and group-specific information from the study population and prove its equivalence to a multi-task Gaussian process (GP) with coregionalization kernel a priori. Compared to multi-task GP, CMDE efficiently handles high-dimensional and multi-modal covariates and provides pointwise uncertainty estimates of causal effects. We evaluate our method across various types of datasets and tasks and find that CMDE outperforms state-of-the-art methods on a majority of these tasks.
comment: 17 pages, 6 figures, 3 tables
♻ ☆ The Solvability of Interpretability Evaluation Metrics EACL 2023
Feature attribution methods are popular for explaining neural network predictions, and they are often evaluated on metrics such as comprehensiveness and sufficiency. In this paper, we highlight an intriguing property of these metrics: their solvability. Concretely, we can define the problem of optimizing an explanation for a metric, which can be solved by beam search. This observation leads to the obvious yet unaddressed question: why do we use explainers (e.g., LIME) not based on solving the target metric, if the metric value represents explanation quality? We present a series of investigations showing strong performance of this beam search explainer and discuss its broader implication: a definition-evaluation duality of interpretability concepts. We implement the explainer and release the Python solvex package for models of text, image and tabular domains.
comment: EACL 2023 (Findings). Project website at https://yilunzhou.github.io/solvability/
♻ ☆ Online Verification of Deep Neural Networks under Domain Shift or Network Updates
Although neural networks are widely used, it remains challenging to formally verify the safety and robustness of neural networks in real-world applications. Existing methods are designed to verify the network before deployment, which are limited to relatively simple specifications and fixed networks. These methods are not ready to be applied to real-world problems with complex and/or dynamically changing specifications and networks. To effectively handle such problems, verification needs to be performed online when these changes take place. However, it is still challenging to run existing verification algorithms online. Our key insight is that we can leverage the temporal dependencies of these changes to accelerate the verification process. This paper establishes a novel framework for scalable online verification to solve real-world verification problems with dynamically changing specifications and/or networks. We propose three types of acceleration algorithms: Branch Management to reduce repetitive computation, Perturbation Tolerance to tolerate changes, and Incremental Computation to reuse previous results. Experiment results show that our algorithms achieve up to $100\times$ acceleration, and thus show a promising way to extend neural network verification to real-world applications.
♻ ☆ Relative Behavioral Attributes: Filling the Gap between Symbolic Goal Specification and Reward Learning from Human Preferences ICLR 2023
Generating complex behaviors that satisfy the preferences of non-expert users is a crucial requirement on AI agents. Interactive reward learning from trajectory comparisons is one way to allow non-expert users to convey complex objectives by expressing preferences over short clips of agent behaviors. Even though this parametric method can encode complex tacit knowledge present in the underlying tasks, it implicitly assumes that the human is unable to provide richer feedback than binary preference labels, leading to intolerably high feedback complexity and poor user experience. While providing a detailed symbolic closed-form specification of the objectives might be tempting, it is not always feasible even for an expert user. However, in most cases, humans are aware of how the agent should change its behavior along meaningful axes to fulfill their underlying purpose, even if they are not able to fully specify task objectives symbolically. Using this as motivation, we introduce the notion of Relative Behavioral Attributes, which allows the users to tweak the agent behavior through symbolic concepts (e.g., increasing the softness or speed of agents' movement). We propose two practical methods that can learn to model any kind of behavioral attributes from ordered behavior clips. We demonstrate the effectiveness of our methods on four tasks with nine different behavioral attributes, showing that once the attributes are learned, end users can produce desirable agent behaviors relatively effortlessly, by providing feedback just around ten times. This is over an order of magnitude less than that required by the popular learning-from-human-preferences baselines. The supplementary video and source code are available at: https://guansuns.github.io/pages/rba.
comment: Published as a conference paper at ICLR 2023
Multimedia 2
♻ ☆ M3FAS: An Accurate and Robust MultiModal Mobile Face Anti-Spoofing System
Face presentation attacks (FPA), also known as face spoofing, have brought increasing concerns to the public through various malicious applications, such as financial fraud and privacy leakage. Therefore, safeguarding face recognition systems against FPA is of utmost importance. Although existing learning-based face anti-spoofing (FAS) models can achieve outstanding detection performance, they lack generalization capability and suffer significant performance drops in unforeseen environments. Many methodologies seek to use auxiliary modality data (e.g., depth and infrared maps) during the presentation attack detection (PAD) to address this limitation. However, these methods can be limited since (1) they require specific sensors such as depth and infrared cameras for data capture, which are rarely available on commodity mobile devices, and (2) they cannot work properly in practical scenarios when either modality is missing or of poor quality. In this paper, we devise an accurate and robust MultiModal Mobile Face Anti-Spoofing system named M3FAS to overcome the issues above. The innovation of this work mainly lies in the following aspects: (1) To achieve robust PAD, our system combines visual and auditory modalities using three pervasively available sensors: camera, speaker, and microphone; (2) We design a novel two-branch neural network with three hierarchical feature aggregation modules to perform cross-modal feature fusion; (3). We propose a multi-head training strategy. The model outputs three predictions from the vision, acoustic, and fusion heads, enabling a more flexible PAD. Extensive experiments have demonstrated the accuracy, robustness, and flexibility of M3FAS under various challenging experimental settings.
♻ ☆ Logically at Factify 2: A Multi-Modal Fact Checking System Based on Evidence Retrieval techniques and Transformer Encoder Architecture AAAI'23
In this paper, we present the Logically submissions to De-Factify 2 challenge (DE-FACTIFY 2023) on the task 1 of Multi-Modal Fact Checking. We describes our submissions to this challenge including explored evidence retrieval and selection techniques, pre-trained cross-modal and unimodal models, and a cross-modal veracity model based on the well established Transformer Encoder (TE) architecture which is heavily relies on the concept of self-attention. Exploratory analysis is also conducted on this Factify 2 data set that uncovers the salient multi-modal patterns and hypothesis motivating the architecture proposed in this work. A series of preliminary experiments were done to investigate and benchmarking different pre-trained embedding models, evidence retrieval settings and thresholds. The final system, a standard two-stage evidence based veracity detection system, yields weighted avg. 0.79 on both val set and final blind test set on the task 1, which achieves 3rd place with a small margin to the top performing system on the leaderboard among 9 participants.
comment: Accepted in AAAI'23: Second Workshop on Multimodal Fact-Checking and Hate Speech Detection, February 2023, Washington, DC, USA
Computation and Language 31
☆ $IC^3$: Image Captioning by Committee Consensus
If you ask a human to describe an image, they might do so in a thousand different ways. Traditionally, image captioning models are trained to approximate the reference distribution of image captions, however, doing so encourages captions that are viewpoint-impoverished. Such captions often focus on only a subset of the possible details, while ignoring potentially useful information in the scene. In this work, we introduce a simple, yet novel, method: "Image Captioning by Committee Consensus" ($IC^3$), designed to generate a single caption that captures high-level details from several viewpoints. Notably, humans rate captions produced by $IC^3$ at least as helpful as baseline SOTA models more than two thirds of the time, and $IC^3$ captions can improve the performance of SOTA automated recall systems by up to 84%, indicating significant material improvements over existing SOTA approaches for visual description. Our code is publicly available at https://github.com/DavidMChan/caption-by-committee
☆ Accelerating Large Language Model Decoding with Speculative Sampling
We present speculative sampling, an algorithm for accelerating transformer decoding by enabling the generation of multiple tokens from each transformer call. Our algorithm relies on the observation that the latency of parallel scoring of short continuations, generated by a faster but less powerful draft model, is comparable to that of sampling a single token from the larger target model. This is combined with a novel modified rejection sampling scheme which preserves the distribution of the target model within hardware numerics. We benchmark speculative sampling with Chinchilla, a 70 billion parameter language model, achieving a 2-2.5x decoding speedup in a distributed setup, without compromising the sample quality or making modifications to the model itself.
☆ Double Permutation Equivariance for Knowledge Graph Completion
This work provides a formalization of Knowledge Graphs (KGs) as a new class of graphs that we denote doubly exchangeable attributed graphs, where node and pairwise (joint 2-node) representations must be equivariant to permutations of both node ids and edge (& node) attributes (relations & node features). Double-permutation equivariant KG representations open a new research direction in KGs. We show that this equivariance imposes a structural representation of relations that allows neural networks to perform complex logical reasoning tasks in KGs. Finally, we introduce a general blueprint for such equivariant representations and test a simple GNN-based double-permutation equivariant neural architecture that achieve 100% Hits@10 test accuracy in both the WN18RRv1 and NELL995v1 inductive KG completion tasks, and can accurately perform logical reasoning tasks that no existing methods can perform, to the best of our knowledge.
☆ What Language Reveals about Perception: Distilling Psychophysical Knowledge from Large Language Models
Understanding the extent to which the perceptual world can be recovered from language is a fundamental problem in cognitive science. We reformulate this problem as that of distilling psychophysical information from text and show how this can be done by combining large language models (LLMs) with a classic psychophysical method based on similarity judgments. Specifically, we use the prompt auto-completion functionality of GPT3, a state-of-the-art LLM, to elicit similarity scores between stimuli and then apply multidimensional scaling to uncover their underlying psychological space. We test our approach on six perceptual domains and show that the elicited judgments strongly correlate with human data and successfully recover well-known psychophysical structures such as the color wheel and pitch spiral. We also explore meaningful divergences between LLM and human representations. Our work showcases how combining state-of-the-art machine models with well-known cognitive paradigms can shed new light on fundamental questions in perception and language research.
comment: 7 pages, 5 figures
☆ Combining Deep Neural Reranking and Unsupervised Extraction for Multi-Query Focused Summarization
The CrisisFACTS Track aims to tackle challenges such as multi-stream fact-finding in the domain of event tracking; participants' systems extract important facts from several disaster-related events while incorporating the temporal order. We propose a combination of retrieval, reranking, and the well-known Integer Linear Programming (ILP) and Maximal Marginal Relevance (MMR) frameworks. In the former two modules, we explore various methods including an entity-based baseline, pre-trained and fine-tuned Question Answering systems, and ColBERT. We then use the latter module as an extractive summarization component by taking diversity and novelty criteria into account. The automatic scoring runs show strong results across the evaluation setups but also reveal shortcomings and challenges.
comment: CrisisFACTS (TREC 2022)
☆ New Linear-time Algorithm for SubTree Kernel Computation based on Root-Weighted Tree Automata
Tree kernels have been proposed to be used in many areas as the automatic learning of natural language applications. In this paper, we propose a new linear time algorithm based on the concept of weighted tree automata for SubTree kernel computation. First, we introduce a new class of weighted tree automata, called Root-Weighted Tree Automata, and their associated formal tree series. Then we define, from this class, the SubTree automata that represent compact computational models for finite tree languages. This allows us to design a theoretically guaranteed linear-time algorithm for computing the SubTree Kernel based on weighted tree automata intersection. The key idea behind the proposed algorithm is to replace DAG reduction and nodes sorting steps used in previous approaches by states equivalence classes computation allowed in the weighted tree automata approach. Our approach has three major advantages: it is output-sensitive, it is free sensitive from the tree types (ordered trees versus unordered trees), and it is well adapted to any incremental tree kernel based learning methods. Finally, we conduct a variety of comparative experiments on a wide range of synthetic tree languages datasets adapted for a deep algorithm analysis. The obtained results show that the proposed algorithm outperforms state-of-the-art methods.
☆ Semantic Coherence Markers for the Early Diagnosis of the Alzheimer Disease
In this work we explore how language models can be employed to analyze language and discriminate between mentally impaired and healthy subjects through the perplexity metric. Perplexity was originally conceived as an information-theoretic measure to assess how much a given language model is suited to predict a text sequence or, equivalently, how much a word sequence fits into a specific language model. We carried out an extensive experimentation with the publicly available data, and employed language models as diverse as N-grams, from 2-grams to 5-grams, and GPT-2, a transformer-based language model. We investigated whether perplexity scores may be used to discriminate between the transcripts of healthy subjects and subjects suffering from Alzheimer Disease (AD). Our best performing models achieved full accuracy and F-score (1.00 in both precision/specificity and recall/sensitivity) in categorizing subjects from both the AD class and control subjects. These results suggest that perplexity can be a valuable analytical metrics with potential application to supporting early diagnosis of symptoms of mental disorders.
comment: This paper is the (significantly) abridged version of the article "Semantic coherence markers: The contribution of perplexity metrics" (https://doi.org/10.1016/j.artmed.2022.102393), which also contains references to employed data and to the implementation of the described work
☆ Predefined domain specific embeddings of food concepts and recipes: A case study on heterogeneous recipe datasets
Although recipe data are very easy to come by nowadays, it is really hard to find a complete recipe dataset - with a list of ingredients, nutrient values per ingredient, and per recipe, allergens, etc. Recipe datasets are usually collected from social media websites where users post and publish recipes. Usually written with little to no structure, using both standardized and non-standardized units of measurement. We collect six different recipe datasets, publicly available, in different formats, and some including data in different languages. Bringing all of these datasets to the needed format for applying a machine learning (ML) pipeline for nutrient prediction [1], [2], includes data normalization using dictionary-based named entity recognition (NER), rule-based NER, as well as conversions using external domain-specific resources. From the list of ingredients, domain-specific embeddings are created using the same embedding space for all recipes - one ingredient dataset is generated. The result from this normalization process is two corpora - one with predefined ingredient embeddings and one with predefined recipe embeddings. On all six recipe datasets, the ML pipeline is evaluated. The results from this use case also confirm that the embeddings merged using the domain heuristic yield better results than the baselines.
comment: 10 pages, 7 tables
☆ Curriculum-guided Abstractive Summarization for Mental Health Online Posts
Automatically generating short summaries from users' online mental health posts could save counselors' reading time and reduce their fatigue so that they can provide timely responses to those seeking help for improving their mental state. Recent Transformers-based summarization models have presented a promising approach to abstractive summarization. They go beyond sentence selection and extractive strategies to deal with more complicated tasks such as novel word generation and sentence paraphrasing. Nonetheless, these models have a prominent shortcoming; their training strategy is not quite efficient, which restricts the model's performance. In this paper, we include a curriculum learning approach to reweigh the training samples, bringing about an efficient learning procedure. We apply our model on extreme summarization dataset of MentSum posts -- a dataset of mental health related posts from Reddit social media. Compared to the state-of-the-art model, our proposed method makes substantial gains in terms of Rouge and Bertscore evaluation metrics, yielding 3.5% (Rouge-1), 10.4% (Rouge-2), and 4.7% (Rouge-L), 1.5% (Bertscore) relative improvements.
comment: 4 pages, short paper, accepted to The 13th International Workshop on Health Text Mining and Information Analysis (LOUHI 2022)
☆ TransFool: An Adversarial Attack against Neural Machine Translation Models
Deep neural networks have been shown to be vulnerable to small perturbations of their inputs, known as adversarial attacks. In this paper, we investigate the vulnerability of Neural Machine Translation (NMT) models to adversarial attacks and propose a new attack algorithm called TransFool. To fool NMT models, TransFool builds on a multi-term optimization problem and a gradient projection step. By integrating the embedding representation of a language model, we generate fluent adversarial examples in the source language that maintain a high level of semantic similarity with the clean samples. Experimental results demonstrate that, for different translation tasks and NMT architectures, our white-box attack can severely degrade the translation quality while the semantic similarity between the original and the adversarial sentences stays high. Moreover, we show that TransFool is transferable to unknown target models. Finally, based on automatic and human evaluations, TransFool leads to improvement in terms of success rate, semantic similarity, and fluency compared to the existing attacks both in white-box and black-box settings. Thus, TransFool permits us to better characterize the vulnerability of NMT models and outlines the necessity to design strong defense mechanisms and more robust NMT systems for real-life applications.
☆ The Fewer Splits are Better: Deconstructing Readability in Sentence Splitting
In this work, we focus on sentence splitting, a subfield of text simplification, motivated largely by an unproven idea that if you divide a sentence in pieces, it should become easier to understand. Our primary goal in this paper is to find out whether this is true. In particular, we ask, does it matter whether we break a sentence into two or three? We report on our findings based on Amazon Mechanical Turk. More specifically, we introduce a Bayesian modeling framework to further investigate to what degree a particular way of splitting the complex sentence affects readability, along with a number of other parameters adopted from diverse perspectives, including clinical linguistics, and cognitive linguistics. The Bayesian modeling experiment provides clear evidence that bisecting the sentence leads to enhanced readability to a degree greater than what we create by trisection.
☆ Multimodal Chain-of-Thought Reasoning in Language Models
Large language models (LLMs) have shown impressive performance on complex reasoning by leveraging chain-of-thought (CoT) prompting to generate intermediate reasoning chains as the rationale to infer the answer. However, existing CoT studies are mostly isolated in the language modality with LLMs, where LLMs are hard to deploy. To elicit CoT reasoning in multimodality, a possible solution is to fine-tune small language models by fusing the vision and language features to perform CoT reasoning. The key challenge is that those language models tend to generate hallucinated reasoning chains that mislead the answer inference. To mitigate the effect of such mistakes, we propose Multimodal-CoT that incorporates vision features in a decoupled training framework. The framework separates the rationale generation and answer inference into two stages. By incorporating the vision features in both stages, the model is able to generate effective rationales that contribute to answer inference. With Multimodal-CoT, our model under 1 billion parameters outperforms the previous state-of-the-art LLM (GPT-3.5) by 16% (75.17%->91.68%) on the ScienceQA benchmark and even surpasses human performance. Code is publicly available at https://github.com/amazon-science/mm-cot.
☆ History-Aware Hierarchical Transformer for Multi-session Open-domain Dialogue System EMNLP 2022
With the evolution of pre-trained language models, current open-domain dialogue systems have achieved great progress in conducting one-session conversations. In contrast, Multi-Session Conversation (MSC), which consists of multiple sessions over a long term with the same user, is under-investigated. In this paper, we propose History-Aware Hierarchical Transformer (HAHT) for multi-session open-domain dialogue. HAHT maintains a long-term memory of history conversations and utilizes history information to understand current conversation context and generate well-informed and context-relevant responses. Specifically, HAHT first encodes history conversation sessions hierarchically into a history memory. Then, HAHT leverages historical information to facilitate the understanding of the current conversation context by encoding the history memory together with the current context with attention-based mechanisms. Finally, to explicitly utilize historical information, HAHT uses a history-aware response generator that switches between a generic vocabulary and a history-aware vocabulary. Experimental results on a large-scale MSC dataset suggest that the proposed HAHT model consistently outperforms baseline models. Human evaluation results support that HAHT generates more human-like, context-relevant and history-relevant responses than baseline models.
comment: EMNLP 2022(Findings)
☆ Language Quantized AutoEncoders: Towards Unsupervised Text-Image Alignment
Recent progress in scaling up large language models has shown impressive capabilities in performing few-shot learning across a wide range of text-based tasks. However, a key limitation is that these language models fundamentally lack visual perception - a crucial attribute needed to extend these models to be able to interact with the real world and solve vision tasks, such as in visual-question answering and robotics. Prior works have largely connected image to text through pretraining and/or fine-tuning on curated image-text datasets, which can be a costly and expensive process. In order to resolve this limitation, we propose a simple yet effective approach called Language-Quantized AutoEncoder (LQAE), a modification of VQ-VAE that learns to align text-image data in an unsupervised manner by leveraging pretrained language models (e.g., BERT, RoBERTa). Our main idea is to encode image as sequences of text tokens by directly quantizing image embeddings using a pretrained language codebook. We then apply random masking followed by a BERT model, and have the decoder reconstruct the original image from BERT predicted text token embeddings. By doing so, LQAE learns to represent similar images with similar clusters of text tokens, thereby aligning these two modalities without the use of aligned text-image pairs. This enables few-shot image classification with large language models (e.g., GPT-3) as well as linear classification of images based on BERT text features. To the best of our knowledge, our work is the first work that uses unaligned images for multimodal tasks by leveraging the power of pretrained language models.
comment: arXiv admin note: text overlap with arXiv:2106.13884 by other authors
☆ How to choose "Good" Samples for Text Data Augmentation
Deep learning-based text classification models need abundant labeled data to obtain competitive performance. Unfortunately, annotating large-size corpus is time-consuming and laborious. To tackle this, multiple researches try to use data augmentation to expand the corpus size. However, data augmentation may potentially produce some noisy augmented samples. There are currently no works exploring sample selection for augmented samples in nature language processing field. In this paper, we propose a novel self-training selection framework with two selectors to select the high-quality samples from data augmentation. Specifically, we firstly use an entropy-based strategy and the model prediction to select augmented samples. Considering some samples with high quality at the above step may be wrongly filtered, we propose to recall them from two perspectives of word overlap and semantic similarity. Experimental results show the effectiveness and simplicity of our framework.
☆ Using In-Context Learning to Improve Dialogue Safety
While large neural-based conversational models have become increasingly proficient as dialogue agents, recent work has highlighted safety issues with these systems. For example, these systems can be goaded into generating toxic content, which often perpetuates social biases or stereotypes. We investigate a retrieval-based framework for reducing bias and toxicity in responses generated from neural-based chatbots. It uses in-context learning to steer a model towards safer generations. Concretely, to generate a response to an unsafe dialogue context, we retrieve demonstrations of safe model responses to similar dialogue contexts. We find our proposed approach performs competitively with strong baselines which use fine-tuning. For instance, using automatic evaluation, we find our best fine-tuned baseline only generates safe responses to unsafe dialogue contexts from DiaSafety 2.92% more than our approach. Finally, we also propose a straightforward re-ranking procedure which can further improve response safeness.
☆ idT5: Indonesian Version of Multilingual T5 Transformer
Indonesian language is spoken by almost 200 million people and is the 10th most spoken language in the world, but it is under-represented in NLP (Natural Language Processing) research. A sparsity of language resources has hampered previous work on Indonesian. The Transformer is a new architecture rapidly becoming dominant for NLP, surpassing alternatives like convolutional and recurrent neural networks. T5 (Text-to-Text Transfer Transformer) is a Transformer model that converts all text-based language problems to text-to-text format for English. The multilingual variant is mT5 (multilingual T5) which has shown promising results on many NLP tasks across languages. However, the size of this multilingual model is a drawback for its application in real production applications, which sometimes require only one language. In this study, the mT5 model was adapted for only one language, Indonesian, resulting in a pre-trained T5 model that was specific only for Indonesian with a smaller size. For performance comparison, we fine-tuned this model and the mT5 model to the Sentiment Analysis (SA), Question Generation (QG), and Question Answering (QA) tasks with the exact mechanism and dataset. Fine-tuned model based on our model achieved 77.18% accuracy on SA, 8% higher than the mT5-based model, and obtained nearly the same score as the mT5-based model on QG and QA. The results confirm that it is possible to produce a smaller pre-trained model that maintains comparable yields while reducing the model size by up to 58%. In addition, the resulting model requires less memory, loads faster, and inference times faster.
☆ Improving Rare Words Recognition through Homophone Extension and Unified Writing for Low-resource Cantonese Speech Recognition SC
Homophone characters are common in tonal syllable-based languages, such as Mandarin and Cantonese. The data-intensive end-to-end Automatic Speech Recognition (ASR) systems are more likely to mis-recognize homophone characters and rare words under low-resource settings. For the problem of lowresource Cantonese speech recognition, this paper presents a novel homophone extension method to integrate human knowledge of the homophone lexicon into the beam search decoding process with language model re-scoring. Besides, we propose an automatic unified writing method to merge the variants of Cantonese characters and standardize speech annotation guidelines, which enables more efficient utilization of labeled utterances by providing more samples for the merged characters. We empirically show that both homophone extension and unified writing improve the recognition performance significantly on both in-domain and out-of-domain test sets, with an absolute Character Error Rate (CER) decrease of around 5% and 18%.
comment: The 13th International Symposium on Chinese Spoken Language Processing (ISCSLP 2022)
☆ CTE: A Dataset for Contextualized Table Extraction
Relevant information in documents is often summarized in tables, helping the reader to identify useful facts. Most benchmark datasets support either document layout analysis or table understanding, but lack in providing data to apply both tasks in a unified way. We define the task of Contextualized Table Extraction (CTE), which aims to extract and define the structure of tables considering the textual context of the document. The dataset comprises 75k fully annotated pages of scientific papers, including more than 35k tables. Data are gathered from PubMed Central, merging the information provided by annotations in the PubTables-1M and PubLayNet datasets. The dataset can support CTE and adds new classes to the original ones. The generated annotations can be used to develop end-to-end pipelines for various tasks, including document layout analysis, table detection, structure recognition, and functional analysis. We formally define CTE and evaluation metrics, showing which subtasks can be tackled, describing advantages, limitations, and future works of this collection of data. Annotations and code will be accessible a https://github.com/AILab-UniFI/cte-dataset.
☆ Commonsense-Aware Prompting for Controllable Empathetic Dialogue Generation AAAI 2023
Improving the emotional awareness of pre-trained language models is an emerging important problem for dialogue generation tasks. Although prior studies have introduced methods to improve empathetic dialogue generation, few have discussed how to incorporate commonsense knowledge into pre-trained language models for controllable dialogue generation. In this study, we propose a novel framework that improves empathetic dialogue generation using pre-trained language models by 1) incorporating commonsense knowledge through prompt verbalization, and 2) controlling dialogue generation using a strategy-driven future discriminator. We conducted experiments to reveal that both the incorporation of social commonsense knowledge and enforcement of control over generation help to improve generation performance. Finally, we discuss the implications of our study for future research.
comment: Accepted to Workshop on Knowledge Augmented Methods for Natural Language Processing, in conjunction with AAAI 2023
☆ The unreasonable effectiveness of few-shot learning for machine translation
We demonstrate the potential of few-shot translation systems, trained with unpaired language data, for both high and low-resource language pairs. We show that with only 5 examples of high-quality translation data shown at inference, a transformer decoder-only model trained solely with self-supervised learning, is able to match specialized supervised state-of-the-art models as well as more general commercial translation systems. In particular, we outperform the best performing system on the WMT'21 English - Chinese news translation task by only using five examples of English - Chinese parallel data at inference. Moreover, our approach in building these models does not necessitate joint multilingual training or back-translation, is conceptually simple and shows the potential to extend to the multilingual setting. Furthermore, the resulting models are two orders of magnitude smaller than state-of-the-art language models. We then analyze the factors which impact the performance of few-shot translation systems, and highlight that the quality of the few-shot demonstrations heavily determines the quality of the translations generated by our models. Finally, we show that the few-shot paradigm also provides a way to control certain attributes of the translation -- we show that we are able to control for regional varieties and formality using only a five examples at inference, paving the way towards controllable machine translation systems.
☆ Curriculum-Guided Abstractive Summarization
Recent Transformer-based summarization models have provided a promising approach to abstractive summarization. They go beyond sentence selection and extractive strategies to deal with more complicated tasks such as novel word generation and sentence paraphrasing. Nonetheless, these models have two shortcomings: (1) they often perform poorly in content selection, and (2) their training strategy is not quite efficient, which restricts model performance. In this paper, we explore two orthogonal ways to compensate for these pitfalls. First, we augment the Transformer network with a sentence cross-attention module in the decoder, encouraging more abstraction of salient content. Second, we include a curriculum learning approach to reweight the training samples, bringing about an efficient learning procedure. Our second approach to enhance the training strategy of Transformers networks makes stronger gains as compared to the first approach. We apply our model on extreme summarization dataset of Reddit TIFU posts. We further look into three cross-domain summarization datasets (Webis-TLDR-17, CNN/DM, and XSum), measuring the efficacy of curriculum learning when applied in summarization. Moreover, a human evaluation is conducted to show the efficacy of the proposed method in terms of qualitative criteria, namely, fluency, informativeness, and overall quality.
comment: 8 pages, Long paper. arXiv admin note: text overlap with arXiv:2302.00954
☆ Creating a Large Language Model of a Philosopher
Can large language models be trained to produce philosophical texts that are difficult to distinguish from texts produced by human philosophers? To address this question, we fine-tuned OpenAI's GPT-3 with the works of philosopher Daniel C. Dennett as additional training data. To explore the Dennett model, we asked the real Dennett ten philosophical questions and then posed the same questions to the language model, collecting four responses for each question without cherry-picking. We recruited 425 participants to distinguish Dennett's answer from the four machine-generated answers. Experts on Dennett's work (N = 25) succeeded 51% of the time, above the chance rate of 20% but short of our hypothesized rate of 80% correct. For two of the ten questions, the language model produced at least one answer that experts selected more frequently than Dennett's own answer. Philosophy blog readers (N = 302) performed similarly to the experts, while ordinary research participants (N = 98) were near chance distinguishing GPT-3's responses from those of an "actual human philosopher".
comment: Earlier version (Nov 2, 2022) available at http://www.faculty.ucr.edu/~eschwitz/SchwitzPapers/GPT-3-Dennett-221102.pdf; stable URL for updated versions http://faculty.ucr.edu/~eschwitz/SchwitzAbs/GPT3Dennett.htm
♻ ☆ GausSetExpander: A Simple Approach for Entity Set Expansion
Entity Set Expansion is an important NLP task that aims at expanding a small set of entities into a larger one with items from a large pool of candidates. In this paper, we propose GausSetExpander, an unsupervised approach based on optimal transport techniques. We propose to re-frame the problem as choosing the entity that best completes the seed set. For this, we interpret a set as an elliptical distribution with a centroid which represents the mean and a spread that is represented by the scale parameter. The best entity is the one that increases the spread of the set the least. We demonstrate the validity of our approach by comparing to state-of-the art approaches.
comment: 8 pages
♻ ☆ BridgeTower: Building Bridges Between Encoders in Vision-Language Representation Learning AAAI 2023
Vision-Language (VL) models with the Two-Tower architecture have dominated visual-language representation learning in recent years. Current VL models either use lightweight uni-modal encoders and learn to extract, align and fuse both modalities simultaneously in a deep cross-modal encoder, or feed the last-layer uni-modal representations from the deep pre-trained uni-modal encoders into the top cross-modal encoder. Both approaches potentially restrict vision-language representation learning and limit model performance. In this paper, we propose BridgeTower, which introduces multiple bridge layers that build a connection between the top layers of uni-modal encoders and each layer of the cross-modal encoder. This enables effective bottom-up cross-modal alignment and fusion between visual and textual representations of different semantic levels of pre-trained uni-modal encoders in the cross-modal encoder. Pre-trained with only 4M images, BridgeTower achieves state-of-the-art performance on various downstream vision-language tasks. In particular, on the VQAv2 test-std set, BridgeTower achieves an accuracy of 78.73%, outperforming the previous state-of-the-art model METER by 1.09% with the same pre-training data and almost negligible additional parameters and computational costs. Notably, when further scaling the model, BridgeTower achieves an accuracy of 81.15%, surpassing models that are pre-trained on orders-of-magnitude larger datasets. Code and checkpoints are available at https://github.com/microsoft/BridgeTower.
comment: Accepted by AAAI 2023, Oral
♻ ☆ ESIE-BERT: Enriching Sub-words Information Explicitly with BERT for Joint Intent Classification and SlotFilling
Natural language understanding (NLU) has two core tasks: intent classification and slot filling. The success of pre-training language models resulted in a significant breakthrough in the two tasks. One of the promising solutions called BERT can jointly optimize the two tasks. We note that BERT-based models convert each complex token into multiple sub-tokens by wordpiece algorithm, which generates a mismatch between the lengths of the tokens and the labels. This leads to BERT-based models do not do well in label prediction which limits model performance improvement. Many existing models can be compatible with this issue but some hidden semantic information is discarded in the fine-tuning process. We address the problem by introducing a novel joint method on top of BERT which explicitly models the multiple sub-tokens features after wordpiece tokenization, thereby contributing to the two tasks. Our method can well extract the contextual features from complex tokens by the proposed sub-words attention adapter (SAA), which preserves overall utterance information. Additionally, we propose an intent attention adapter (IAA) to obtain the full sentence features to aid users to predict intent. Experimental results confirm that our proposed model is significantly improved on two public benchmark datasets. In particular, the slot filling F1 score is improved from 96.1 to 98.2 (2.1% absolute) on the Airline Travel Information Systems (ATIS) dataset.
♻ ☆ Knowledge Extraction in Low-Resource Scenarios: Survey and Perspective
Knowledge Extraction (KE), aiming to extract structural information from unstructured texts, often suffers from data scarcity and emerging unseen types, i.e., low-resource scenarios. Many neural approaches to low-resource KE have been widely investigated and achieved impressive performance. In this paper, we present a literature review towards KE in low-resource scenarios, and systematically categorize existing works into three paradigms: (1) exploiting higher-resource data, (2) exploiting stronger models, and (3) exploiting data and models together. In addition, we highlight promising applications and outline some potential directions for future research. We hope that our survey can help both the academic and industrial communities to better understand this field, inspire more ideas, and boost broader applications.
comment: Work in Progress
♻ ☆ Grammar construction methods for extended deterministic expressions
Extended regular expressions with counting and interleaving are widely used in practice. However the related theoretical studies for this kind of expressions currently cannot meet the need of practical work. This paper develops syntax definitions for extended deterministic expressions and their subclasses, hope to completely solve the long-standing problem that there are no syntax definitions for this kind of expressions, which has become an important reason for restricting the use of extended expressions.
comment: in Chinese language
♻ ☆ PiC: A Phrase-in-Context Dataset for Phrase Understanding and Semantic Search EACL 2023
While contextualized word embeddings have been a de-facto standard, learning contextualized phrase embeddings is less explored and being hindered by the lack of a human-annotated benchmark that tests machine understanding of phrase semantics given a context sentence or paragraph (instead of phrases alone). To fill this gap, we propose PiC -- a dataset of ~28K of noun phrases accompanied by their contextual Wikipedia pages and a suite of three tasks for training and evaluating phrase embeddings. Training on PiC improves ranking models' accuracy and remarkably pushes span-selection (SS) models (i.e., predicting the start and end index of the target phrase) near-human accuracy, which is 95% Exact Match (EM) on semantic search given a query phrase and a passage. Interestingly, we find evidence that such impressive performance is because the SS models learn to better capture the common meaning of a phrase regardless of its actual context. SotA models perform poorly in distinguishing two senses of the same phrase in two contexts (~60% EM) and in estimating the similarity between two different phrases in the same context (~70% EM).
comment: Accepted to EACL 2023
♻ ☆ A Survey on Table-and-Text HybridQA: Concepts, Methods, Challenges and Future Directions
Table-and-text hybrid question answering (HybridQA) is a widely used and challenging NLP task commonly applied in the financial and scientific domain. The early research focuses on migrating other QA task methods to HybridQA, while with further research, more and more HybridQA-specific methods have been present. With the rapid development of HybridQA, the systematic survey is still under-explored to summarize the main techniques and advance further research. So we present this work to summarize the current HybridQA benchmarks and methods, then analyze the challenges and future directions of this task. The contributions of this paper can be summarized in three folds: (1) first survey, to our best knowledge, including benchmarks, methods and challenges for HybridQA; (2) systematic investigation with the reasonable comparison of the existing systems to articulate their advantages and shortcomings; (3) detailed analysis of challenges in four important dimensions to shed light on future directions.
comment: 7 pages
♻ ☆ Summarization Programs: Interpretable Abstractive Summarization with Neural Modular Trees ICLR 2023
Current abstractive summarization models either suffer from a lack of clear interpretability or provide incomplete rationales by only highlighting parts of the source document. To this end, we propose the Summarization Program (SP), an interpretable modular framework consisting of an (ordered) list of binary trees, each encoding the step-by-step generative process of an abstractive summary sentence from the source document. A Summarization Program contains one root node per summary sentence, and a distinct tree connects each summary sentence (root node) to the document sentences (leaf nodes) from which it is derived, with the connecting nodes containing intermediate generated sentences. Edges represent different modular operations involved in summarization such as sentence fusion, compression, and paraphrasing. We first propose an efficient best-first search method over neural modules, SP-Search that identifies SPs for human summaries by directly optimizing for ROUGE scores. Next, using these programs as automatic supervision, we propose seq2seq models that generate Summarization Programs, which are then executed to obtain final summaries. We demonstrate that SP-Search effectively represents the generative process behind human summaries using modules that are typically faithful to their intended behavior. We also conduct a simulation study to show that Summarization Programs improve the interpretability of summarization models by allowing humans to better simulate model reasoning. Summarization Programs constitute a promising step toward interpretable and modular abstractive summarization, a complex task previously addressed primarily through blackbox end-to-end neural systems. Supporting code available at https://github.com/swarnaHub/SummarizationPrograms
comment: ICLR 2023
Computer Vision and Pattern Recognition 93
☆ STEPS: Joint Self-supervised Nighttime Image Enhancement and Depth Estimation ICRA 2023
Self-supervised depth estimation draws a lot of attention recently as it can promote the 3D sensing capabilities of self-driving vehicles. However, it intrinsically relies upon the photometric consistency assumption, which hardly holds during nighttime. Although various supervised nighttime image enhancement methods have been proposed, their generalization performance in challenging driving scenarios is not satisfactory. To this end, we propose the first method that jointly learns a nighttime image enhancer and a depth estimator, without using ground truth for either task. Our method tightly entangles two self-supervised tasks using a newly proposed uncertain pixel masking strategy. This strategy originates from the observation that nighttime images not only suffer from underexposed regions but also from overexposed regions. By fitting a bridge-shaped curve to the illumination map distribution, both regions are suppressed and two tasks are bridged naturally. We benchmark the method on two established datasets: nuScenes and RobotCar and demonstrate state-of-the-art performance on both of them. Detailed ablations also reveal the mechanism of our proposal. Last but not least, to mitigate the problem of sparse ground truth of existing datasets, we provide a new photo-realistically enhanced nighttime dataset based upon CARLA. It brings meaningful new challenges to the community. Codes, data, and models are available at https://github.com/ucaszyp/STEPS.
comment: Accepted by ICRA 2023, Code: https://github.com/ucaszyp/STEPS
☆ Bayesian Metric Learning for Uncertainty Quantification in Image Retrieval
We propose the first Bayesian encoder for metric learning. Rather than relying on neural amortization as done in prior works, we learn a distribution over the network weights with the Laplace Approximation. We actualize this by first proving that the contrastive loss is a valid log-posterior. We then propose three methods that ensure a positive definite Hessian. Lastly, we present a novel decomposition of the Generalized Gauss-Newton approximation. Empirically, we show that our Laplacian Metric Learner (LAM) estimates well-calibrated uncertainties, reliably detects out-of-distribution examples, and yields state-of-the-art predictive performance.
comment: Code: https://github.com/FrederikWarburg/bayesian-metric-learning
☆ SceneDreamer: Unbounded 3D Scene Generation from 2D Image Collections
In this work, we present SceneDreamer, an unconditional generative model for unbounded 3D scenes, which synthesizes large-scale 3D landscapes from random noises. Our framework is learned from in-the-wild 2D image collections only, without any 3D annotations. At the core of SceneDreamer is a principled learning paradigm comprising 1) an efficient yet expressive 3D scene representation, 2) a generative scene parameterization, and 3) an effective renderer that can leverage the knowledge from 2D images. Our framework starts from an efficient bird's-eye-view (BEV) representation generated from simplex noise, which consists of a height field and a semantic field. The height field represents the surface elevation of 3D scenes, while the semantic field provides detailed scene semantics. This BEV scene representation enables 1) representing a 3D scene with quadratic complexity, 2) disentangled geometry and semantics, and 3) efficient training. Furthermore, we propose a novel generative neural hash grid to parameterize the latent space given 3D positions and the scene semantics, which aims to encode generalizable features across scenes. Lastly, a neural volumetric renderer, learned from 2D image collections through adversarial training, is employed to produce photorealistic images. Extensive experiments demonstrate the effectiveness of SceneDreamer and superiority over state-of-the-art methods in generating vivid yet diverse unbounded 3D worlds.
comment: Project Page https://scene-dreamer.github.io/
☆ Dreamix: Video Diffusion Models are General Video Editors
Text-driven image and video diffusion models have recently achieved unprecedented generation realism. While diffusion models have been successfully applied for image editing, very few works have done so for video editing. We present the first diffusion-based method that is able to perform text-based motion and appearance editing of general videos. Our approach uses a video diffusion model to combine, at inference time, the low-resolution spatio-temporal information from the original video with new, high resolution information that it synthesized to align with the guiding text prompt. As obtaining high-fidelity to the original video requires retaining some of its high-resolution information, we add a preliminary stage of finetuning the model on the original video, significantly boosting fidelity. We propose to improve motion editability by a new, mixed objective that jointly finetunes with full temporal attention and with temporal attention masking. We further introduce a new framework for image animation. We first transform the image into a coarse video by simple image processing operations such as replication and perspective geometric projections, and then use our general video editor to animate it. As a further application, we can use our method for subject-driven video generation. Extensive qualitative and numerical experiments showcase the remarkable editing ability of our method and establish its superior performance compared to baseline methods.
☆ $IC^3$: Image Captioning by Committee Consensus
If you ask a human to describe an image, they might do so in a thousand different ways. Traditionally, image captioning models are trained to approximate the reference distribution of image captions, however, doing so encourages captions that are viewpoint-impoverished. Such captions often focus on only a subset of the possible details, while ignoring potentially useful information in the scene. In this work, we introduce a simple, yet novel, method: "Image Captioning by Committee Consensus" ($IC^3$), designed to generate a single caption that captures high-level details from several viewpoints. Notably, humans rate captions produced by $IC^3$ at least as helpful as baseline SOTA models more than two thirds of the time, and $IC^3$ captions can improve the performance of SOTA automated recall systems by up to 84%, indicating significant material improvements over existing SOTA approaches for visual description. Our code is publicly available at https://github.com/DavidMChan/caption-by-committee
☆ Dual PatchNorm
We propose Dual PatchNorm: two Layer Normalization layers (LayerNorms), before and after the patch embedding layer in Vision Transformers. We demonstrate that Dual PatchNorm outperforms the result of exhaustive search for alternative LayerNorm placement strategies in the Transformer block itself. In our experiments, incorporating this trivial modification, often leads to improved accuracy over well-tuned Vision Transformers and never hurts.
☆ Are Diffusion Models Vulnerable to Membership Inference Attacks?
Diffusion-based generative models have shown great potential for image synthesis, but there is a lack of research on the security and privacy risks they may pose. In this paper, we investigate the vulnerability of diffusion models to Membership Inference Attacks (MIAs), a common privacy concern. Our results indicate that existing MIAs designed for GANs or VAE are largely ineffective on diffusion models, either due to inapplicable scenarios (e.g., requiring the discriminator of GANs) or inappropriate assumptions (e.g., closer distances between synthetic images and member images). To address this gap, we propose Step-wise Error Comparing Membership Inference (SecMI), a black-box MIA that infers memberships by assessing the matching of forward process posterior estimation at each timestep. SecMI follows the common overfitting assumption in MIA where member samples normally have smaller estimation errors, compared with hold-out samples. We consider both the standard diffusion models, e.g., DDPM, and the text-to-image diffusion models, e.g., Stable Diffusion. Experimental results demonstrate that our methods precisely infer the membership with high confidence on both of the two scenarios across six different datasets
☆ Ditto in the House: Building Articulation Models of Indoor Scenes through Interactive Perception ICRA 2023
Virtualizing the physical world into virtual models has been a critical technique for robot navigation and planning in the real world. To foster manipulation with articulated objects in everyday life, this work explores building articulation models of indoor scenes through a robot's purposeful interactions in these scenes. Prior work on articulation reasoning primarily focuses on siloed objects of limited categories. To extend to room-scale environments, the robot has to efficiently and effectively explore a large-scale 3D space, locate articulated objects, and infer their articulations. We introduce an interactive perception approach to this task. Our approach, named Ditto in the House, discovers possible articulated objects through affordance prediction, interacts with these objects to produce articulated motions, and infers the articulation properties from the visual observations before and after each interaction. It tightly couples affordance prediction and articulation inference to improve both tasks. We demonstrate the effectiveness of our approach in both simulation and real-world scenes. Code and additional results are available at https://ut-austin-rpl.github.io/HouseDitto/
comment: ICRA 2023. Code and additional results are available at https://ut-austin-rpl.github.io/HouseDitto/
☆ Multi-scale Feature Alignment for Continual Learning of Unlabeled Domains
Methods for unsupervised domain adaptation (UDA) help to improve the performance of deep neural networks on unseen domains without any labeled data. Especially in medical disciplines such as histopathology, this is crucial since large datasets with detailed annotations are scarce. While the majority of existing UDA methods focus on the adaptation from a labeled source to a single unlabeled target domain, many real-world applications with a long life cycle involve more than one target domain. Thus, the ability to sequentially adapt to multiple target domains becomes essential. In settings where the data from previously seen domains cannot be stored, e.g., due to data protection regulations, the above becomes a challenging continual learning problem. To this end, we propose to use generative feature-driven image replay in conjunction with a dual-purpose discriminator that not only enables the generation of images with realistic features for replay, but also promotes feature alignment during domain adaptation. We evaluate our approach extensively on a sequence of three histopathological datasets for tissue-type classification, achieving state-of-the-art results. We present detailed ablation experiments studying our proposed method components and demonstrate a possible use-case of our continual UDA method for an unsupervised patch-based segmentation task given high-resolution tissue images.
☆ Human not in the loop: objective sample difficulty measures for Curriculum Learning
Curriculum learning is a learning method that trains models in a meaningful order from easier to harder samples. A key here is to devise automatic and objective difficulty measures of samples. In the medical domain, previous work applied domain knowledge from human experts to qualitatively assess classification difficulty of medical images to guide curriculum learning, which requires extra annotation efforts, relies on subjective human experience, and may introduce bias. In this work, we propose a new automated curriculum learning technique using the variance of gradients (VoG) to compute an objective difficulty measure of samples and evaluated its effects on elbow fracture classification from X-ray images. Specifically, we used VoG as a metric to rank each sample in terms of the classification difficulty, where high VoG scores indicate more difficult cases for classification, to guide the curriculum training process We compared the proposed technique to a baseline (without curriculum learning), a previous method that used human annotations on classification difficulty, and anti-curriculum learning. Our experiment results showed comparable and higher performance for the binary and multi-class bone fracture classification tasks.
comment: ISBI 2023
☆ Factor Fields: A Unified Framework for Neural Fields and Beyond
We present Factor Fields, a novel framework for modeling and representing signals. Factor Fields decomposes a signal into a product of factors, each of which is represented by a neural or regular field representation operating on a coordinate transformed input signal. We show that this decomposition yields a unified framework that generalizes several recent signal representations including NeRF, PlenOxels, EG3D, Instant-NGP, and TensoRF. Moreover, the framework allows for the creation of powerful new signal representations, such as the Coefficient-Basis Factorization (CoBaFa) which we propose in this paper. As evidenced by our experiments, CoBaFa leads to improvements over previous fast reconstruction methods in terms of the three critical goals in neural signal representation: approximation quality, compactness and efficiency. Experimentally, we demonstrate that our representation achieves better image approximation quality on 2D image regression tasks, higher geometric quality when reconstructing 3D signed distance fields and higher compactness for radiance field reconstruction tasks compared to previous fast reconstruction methods. Besides, our CoBaFa representation enables generalization by sharing the basis across signals during training, enabling generalization tasks such as image regression with sparse observations and few-shot radiance field reconstruction.
comment: 13 pages, 7 figures
☆ Boosting Low-Data Instance Segmentation by Unsupervised Pre-training with Saliency Prompt
Recently, inspired by DETR variants, query-based end-to-end instance segmentation (QEIS) methods have outperformed CNN-based models on large-scale datasets. Yet they would lose efficacy when only a small amount of training data is available since it's hard for the crucial queries/kernels to learn localization and shape priors. To this end, this work offers a novel unsupervised pre-training solution for low-data regimes. Inspired by the recent success of the Prompting technique, we introduce a new pre-training method that boosts QEIS models by giving Saliency Prompt for queries/kernels. Our method contains three parts: 1) Saliency Masks Proposal is responsible for generating pseudo masks from unlabeled images based on the saliency mechanism. 2) Prompt-Kernel Matching transfers pseudo masks into prompts and injects the corresponding localization and shape priors to the best-matched kernels. 3) Kernel Supervision is applied to supply supervision at the kernel level for robust learning. From a practical perspective, our pre-training method helps QEIS models achieve a similar convergence speed and comparable performance with CNN-based models in low-data regimes. Experimental results show that our method significantly boosts several QEIS models on three datasets. Code will be made available.
☆ Get3DHuman: Lifting StyleGAN-Human into a 3D Generative Model using Pixel-aligned Reconstruction Priors
Fast generation of high-quality 3D digital humans is important to a vast number of applications ranging from entertainment to professional concerns. Recent advances in differentiable rendering have enabled the training of 3D generative models without requiring 3D ground truths. However, the quality of the generated 3D humans still has much room to improve in terms of both fidelity and diversity. In this paper, we present Get3DHuman, a novel 3D human framework that can significantly boost the realism and diversity of the generated outcomes by only using a limited budget of 3D ground-truth data. Our key observation is that the 3D generator can profit from human-related priors learned through 2D human generators and 3D reconstructors. Specifically, we bridge the latent space of Get3DHuman with that of StyleGAN-Human via a specially-designed prior network, where the input latent code is mapped to the shape and texture feature volumes spanned by the pixel-aligned 3D reconstructor. The outcomes of the prior network are then leveraged as the supervisory signals for the main generator network. To ensure effective training, we further propose three tailored losses applied to the generated feature volumes and the intermediate feature maps. Extensive experiments demonstrate that Get3DHuman greatly outperforms the other state-of-the-art approaches and can support a wide range of applications including shape interpolation, shape re-texturing, and single-view reconstruction through latent inversion.
☆ UW-CVGAN: UnderWater Image Enhancement with Capsules Vectors Quantization
The degradation in the underwater images is due to wavelength-dependent light attenuation, scattering, and to the diversity of the water types in which they are captured. Deep neural networks take a step in this field, providing autonomous models able to achieve the enhancement of underwater images. We introduce Underwater Capsules Vectors GAN UWCVGAN based on the discrete features quantization paradigm from VQGAN for this task. The proposed UWCVGAN combines an encoding network, which compresses the image into its latent representation, with a decoding network, able to reconstruct the enhancement of the image from the only latent representation. In contrast with VQGAN, UWCVGAN achieves feature quantization by exploiting the clusterization ability of capsule layer, making the model completely trainable and easier to manage. The model obtains enhanced underwater images with high quality and fine details. Moreover, the trained encoder is independent of the decoder giving the possibility to be embedded onto the collector as compressing algorithm to reduce the memory space required for the images, of factor $3\times$. \myUWCVGAN{ }is validated with quantitative and qualitative analysis on benchmark datasets, and we present metrics results compared with the state of the art.
☆ SceneScape: Text-Driven Consistent Scene Generation
We propose a method for text-driven perpetual view generation -- synthesizing long videos of arbitrary scenes solely from an input text describing the scene and camera poses. We introduce a novel framework that generates such videos in an online fashion by combining the generative power of a pre-trained text-to-image model with the geometric priors learned by a pre-trained monocular depth prediction model. To achieve 3D consistency, i.e., generating videos that depict geometrically-plausible scenes, we deploy an online test-time training to encourage the predicted depth map of the current frame to be geometrically consistent with the synthesized scene; the depth maps are used to construct a unified mesh representation of the scene, which is updated throughout the generation and is used for rendering. In contrast to previous works, which are applicable only for limited domains (e.g., landscapes), our framework generates diverse scenes, such as walkthroughs in spaceships, caves, or ice castles. Project page: https://scenescape.github.io/
comment: Project page: https://scenescape.github.io/
☆ A Simple Baseline for Direct 2D Multi-Person Head Pose Estimation with Full-range Angles
Existing head pose estimation (HPE) mainly focuses on single person with pre-detected frontal heads, which limits their applications in real complex scenarios with multi-persons. We argue that these single HPE methods are fragile and inefficient for Multi-Person Head Pose Estimation (MPHPE) since they rely on the separately trained face detector that cannot generalize well to full viewpoints, especially for heads with invisible face areas. In this paper, we focus on the full-range MPHPE problem, and propose a direct end-to-end simple baseline named DirectMHP. Due to the lack of datasets applicable to the full-range MPHPE, we firstly construct two benchmarks by extracting ground-truth labels for head detection and head orientation from public datasets AGORA and CMU Panoptic. They are rather challenging for having many truncated, occluded, tiny and unevenly illuminated human heads. Then, we design a novel end-to-end trainable one-stage network architecture by joint regressing locations and orientations of multi-head to address the MPHPE problem. Specifically, we regard pose as an auxiliary attribute of the head, and append it after the traditional object prediction. Arbitrary pose representation such as Euler angles is acceptable by this flexible design. Then, we jointly optimize these two tasks by sharing features and utilizing appropriate multiple losses. In this way, our method can implicitly benefit from more surroundings to improve HPE accuracy while maintaining head detection performance. We present comprehensive comparisons with state-of-the-art single HPE methods on public benchmarks, as well as superior baseline results on our constructed MPHPE datasets. Datasets and code are released in https://github.com/hnuzhy/DirectMHP.
☆ GraphReg: Dynamical Point Cloud Registration with Geometry-aware Graph Signal Processing
This study presents a high-accuracy, efficient, and physically induced method for 3D point cloud registration, which is the core of many important 3D vision problems. In contrast to existing physics-based methods that merely consider spatial point information and ignore surface geometry, we explore geometry aware rigid-body dynamics to regulate the particle (point) motion, which results in more precise and robust registration. Our proposed method consists of four major modules. First, we leverage the graph signal processing (GSP) framework to define a new signature, (i.e., point response intensity for each point), by which we succeed in describing the local surface variation, resampling keypoints, and distinguishing different particles. Then, to address the shortcomings of current physics-based approaches that are sensitive to outliers, we accommodate the defined point response intensity to median absolute deviation (MAD) in robust statistics and adopt the X84 principle for adaptive outlier depression, ensuring a robust and stable registration. Subsequently, we propose a novel geometric invariant under rigid transformations to incorporate higher-order features of point clouds, which is further embedded for force modeling to guide the correspondence between pairwise scans credibly. Finally, we introduce an adaptive simulated annealing (ASA) method to search for the global optimum and substantially accelerate the registration process. We perform comprehensive experiments to evaluate the proposed method on various datasets captured from range scanners to LiDAR. Results demonstrate that our proposed method outperforms representative state-of-the-art approaches in terms of accuracy and is more suitable for registering large-scale point clouds. Furthermore, it is considerably faster and more robust than most competitors.
☆ A Survey on Efficient Training of Transformers
Recent advances in Transformers have come with a huge requirement on computing resources, highlighting the importance of developing efficient training techniques to make Transformer training faster, at lower cost, and to higher accuracy by the efficient use of computation and memory resources. This survey provides the first systematic overview of the efficient training of Transformers, covering the recent progress in acceleration arithmetic and hardware, with a focus on the former. We analyze and compare methods that save computation and memory costs for intermediate tensors during training, together with techniques on hardware/algorithm co-design. We finally discuss challenges and promising areas for future research.
comment: A brief review
☆ LesionAid: Vision Transformers-based Skin Lesion Generation and Classification
Skin cancer is one of the most prevalent forms of human cancer. It is recognized mainly visually, beginning with clinical screening and continuing with the dermoscopic examination, histological assessment, and specimen collection. Deep convolutional neural networks (CNNs) perform highly segregated and potentially universal tasks against a classified finegrained object. This research proposes a novel multi-class prediction framework that classifies skin lesions based on ViT and ViTGAN. Vision transformers-based GANs (Generative Adversarial Networks) are utilized to tackle the class imbalance. The framework consists of four main phases: ViTGANs, Image processing, and explainable AI. Phase 1 consists of generating synthetic images to balance all the classes in the dataset. Phase 2 consists of applying different data augmentation techniques and morphological operations to increase the size of the data. Phases 3 & 4 involve developing a ViT model for edge computing systems that can identify patterns and categorize skin lesions from the user's skin visible in the image. In phase 3, after classifying the lesions into the desired class with ViT, we will use explainable AI (XAI) that leads to more explainable results (using activation maps, etc.) while ensuring high predictive accuracy. Real-time images of skin diseases can capture by a doctor or a patient using the camera of a mobile application to perform an early examination and determine the cause of the skin lesion. The whole framework is compared with the existing frameworks for skin lesion detection.
☆ Curriculum Learning for ab initio Deep Learned Refractive Optics
Deep lens optimization has recently emerged as a new paradigm for designing computational imaging systems, however it has been limited to either simple optical systems consisting of a single DOE or metalens, or the fine-tuning of compound lenses from good initial designs. Here we present a deep lens design method based on curriculum learning, which is able to learn optical designs of compound lenses ab initio from randomly initialized surfaces, therefore overcoming the need for a good initial design. We demonstrate this approach with the fully-automatic design of an extended depth-of-field computational camera in a cellphone-style form factor, highly aspherical surfaces, and a short back focal length.
comment: Automatically design computational lenses from scratch with differentiable ray tracing
☆ NDJIR: Neural Direct and Joint Inverse Rendering for Geometry, Lights, and Materials of Real Object
The goal of inverse rendering is to decompose geometry, lights, and materials given pose multi-view images. To achieve this goal, we propose neural direct and joint inverse rendering, NDJIR. Different from prior works which relies on some approximations of the rendering equation, NDJIR directly addresses the integrals in the rendering equation and jointly decomposes geometry: signed distance function, lights: environment and implicit lights, materials: base color, roughness, specular reflectance using the powerful and flexible volume rendering framework, voxel grid feature, and Bayesian prior. Our method directly uses the physically-based rendering, so we can seamlessly export an extracted mesh with materials to DCC tools and show material conversion examples. We perform intensive experiments to show that our proposed method can decompose semantically well for real object in photogrammetric setting and what factors contribute towards accurate inverse rendering.
comment: 26 pages
☆ Physics Constrained Motion Prediction with Uncertainty Quantification
Predicting the motion of dynamic agents is a critical task for guaranteeing the safety of autonomous systems. A particular challenge is that motion prediction algorithms should obey dynamics constraints and quantify prediction uncertainty as a measure of confidence. We present a physics-constrained approach for motion prediction which uses a surrogate dynamical model to ensure that predicted trajectories are dynamically feasible. We propose a two-step integration consisting of intent and trajectory prediction subject to dynamics constraints. We also construct prediction regions that quantify uncertainty and are tailored for autonomous driving by using conformal prediction, a popular statistical tool. Physics Constrained Motion Prediction achieves a 41% better ADE, 56% better FDE, and 19% better IoU over a baseline in experiments using an autonomous racing dataset.
comment: Submitted to IV 2023
☆ IKOL: Inverse kinematics optimization layer for 3D human pose and shape estimation via Gauss-Newton differentiation AAAI 2023
This paper presents an inverse kinematic optimization layer (IKOL) for 3D human pose and shape estimation that leverages the strength of both optimization- and regression-based methods within an end-to-end framework. IKOL involves a nonconvex optimization that establishes an implicit mapping from an image's 3D keypoints and body shapes to the relative body-part rotations. The 3D keypoints and the body shapes are the inputs and the relative body-part rotations are the solutions. However, this procedure is implicit and hard to make differentiable. So, to overcome this issue, we designed a Gauss-Newton differentiation (GN-Diff) procedure to differentiate IKOL. GN-Diff iteratively linearizes the nonconvex objective function to obtain Gauss-Newton directions with closed form solutions. Then, an automatic differentiation procedure is directly applied to generate a Jacobian matrix for end-to-end training. Notably, the GN-Diff procedure works fast because it does not rely on a time-consuming implicit differentiation procedure. The twist rotation and shape parameters are learned from the neural networks and, as a result, IKOL has a much lower computational overhead than most existing optimization-based methods. Additionally, compared to existing regression-based methods, IKOL provides a more accurate mesh-image correspondence. This is because it iteratively reduces the distance between the keypoints and also enhances the reliability of the pose structures. Extensive experiments demonstrate the superiority of our proposed framework over a wide range of 3D human pose and shape estimation methods.
comment: Accepted by AAAI 2023
☆ Beyond Pretrained Features: Noisy Image Modeling Provides Adversarial Defense
Masked Image Modeling (MIM) has been a prevailing framework for self-supervised visual representation learning. Within the pretraining-finetuning paradigm, the MIM framework trains an encoder by reconstructing masked image patches with the help of a decoder which would be abandoned when the encoder is used for finetuning. Despite its state-of-the-art performance on clean images, MIM models are vulnerable to adversarial attacks, limiting its real-world application, and few studies have focused on this issue. In this paper, we have discovered that noisy image modeling (NIM), a variant of MIM that uses denoising as the pre-text task, provides not only good pretrained visual features, but also effective adversarial defense for downstream models. To achieve a better accuracy-robustness trade-off, we further propose to sample the hyperparameter that controls the reconstruction difficulty from random distributions instead of setting it globally, and fine-tune downstream networks with denoised images. Experimental results demonstrate that our pre-trained denoising autoencoders are effective against different white-box, gray-box, and black-box attacks without being trained with adversarial images, while not harming the clean accuracy of fine-tuned models. Source code and models will be made available.
☆ Paced-Curriculum Distillation with Prediction and Label Uncertainty for Image Segmentation
Purpose: In curriculum learning, the idea is to train on easier samples first and gradually increase the difficulty, while in self-paced learning, a pacing function defines the speed to adapt the training progress. While both methods heavily rely on the ability to score the difficulty of data samples, an optimal scoring function is still under exploration. Methodology: Distillation is a knowledge transfer approach where a teacher network guides a student network by feeding a sequence of random samples. We argue that guiding student networks with an efficient curriculum strategy can improve model generalization and robustness. For this purpose, we design an uncertainty-based paced curriculum learning in self distillation for medical image segmentation. We fuse the prediction uncertainty and annotation boundary uncertainty to develop a novel paced-curriculum distillation (PCD). We utilize the teacher model to obtain prediction uncertainty and spatially varying label smoothing with Gaussian kernel to generate segmentation boundary uncertainty from the annotation. We also investigate the robustness of our method by applying various types and severity of image perturbation and corruption. Results: The proposed technique is validated on two medical datasets of breast ultrasound image segmentation and robotassisted surgical scene segmentation and achieved significantly better performance in terms of segmentation and robustness. Conclusion: P-CD improves the performance and obtains better generalization and robustness over the dataset shift. While curriculum learning requires extensive tuning of hyper-parameters for pacing function, the level of performance improvement suppresses this limitation.
comment: 15 pages
☆ Real-Time Evaluation in Online Continual Learning: A New Paradigm
Current evaluations of Continual Learning (CL) methods typically assume that there is no constraint on training time and computation. This is an unrealistic assumption for any real-world setting, which motivates us to propose: a practical real-time evaluation of continual learning, in which the stream does not wait for the model to complete training before revealing the next data for predictions. To do this, we evaluate current CL methods with respect to their computational costs. We hypothesize that under this new evaluation paradigm, computationally demanding CL approaches may perform poorly on streams with a varying distribution. We conduct extensive experiments on CLOC, a large-scale dataset containing 39 million time-stamped images with geolocation labels. We show that a simple baseline outperforms state-of-the-art CL methods under this evaluation, questioning the applicability of existing methods in realistic settings. In addition, we explore various CL components commonly used in the literature, including memory sampling strategies and regularization approaches. We find that all considered methods fail to be competitive against our simple baseline. This surprisingly suggests that the majority of existing CL literature is tailored to a specific class of streams that is not practical. We hope that the evaluation we provide will be the first step towards a paradigm shift to consider the computational cost in the development of online continual learning methods.
☆ An Efficient Convex Hull-Based Vehicle Pose Estimation Method for 3D LiDAR
Vehicle pose estimation is essential in the perception technology of autonomous driving. However, due to the different density distributions of the LiDAR point cloud, it is challenging to achieve accurate direction extraction based on 3D LiDAR by using the existing pose estimation methods. In this paper, we proposed a novel convex hull-based vehicle pose estimation method. The extracted 3D cluster is reduced to the convex hull, reducing the computation burden. Then a novel criterion based on the minimum occlusion area is developed for the search-based algorithm, which can achieve accurate pose estimation. The proposed algorithm is validated on the KITTI dataset and a manually labeled dataset acquired at an industrial park. The results show that our proposed method can achieve better accuracy than the three mainstream algorithms while maintaining real-time speed.
☆ FCB-SwinV2 Transformer for Polyp Segmentation
Polyp segmentation within colonoscopy video frames using deep learning models has the potential to automate the workflow of clinicians. This could help improve the early detection rate and characterization of polyps which could progress to colorectal cancer. Recent state-of-the-art deep learning polyp segmentation models have combined the outputs of Fully Convolutional Network architectures and Transformer Network architectures which work in parallel. In this paper we propose modifications to the current state-of-the-art polyp segmentation model FCBFormer. The transformer architecture of the FCBFormer is replaced with a SwinV2 Transformer-UNET and minor changes to the Fully Convolutional Network architecture are made to create the FCB-SwinV2 Transformer. The performance of the FCB-SwinV2 Transformer is evaluated on the popular colonoscopy segmentation bench-marking datasets Kvasir-SEG and CVC-ClinicDB. Generalizability tests are also conducted. The FCB-SwinV2 Transformer is able to consistently achieve higher mDice scores across all tests conducted and therefore represents new state-of-the-art performance. Issues found with how colonoscopy segmentation model performance is evaluated within literature are also re-ported and discussed. One of the most important issues identified is that when evaluating performance on the CVC-ClinicDB dataset it would be preferable to ensure no data leakage from video sequences occurs during the training/validation/test data partition.
☆ Open-Set Multi-Source Multi-Target Domain Adaptation NeurIPS 2021
Single-Source Single-Target Domain Adaptation (1S1T) aims to bridge the gap between a labelled source domain and an unlabelled target domain. Despite 1S1T being a well-researched topic, they are typically not deployed to the real world. Methods like Multi-Source Domain Adaptation and Multi-Target Domain Adaptation have evolved to model real-world problems but still do not generalise well. The fact that most of these methods assume a common label-set between source and target is very restrictive. Recent Open-Set Domain Adaptation methods handle unknown target labels but fail to generalise in multiple domains. To overcome these difficulties, first, we propose a novel generic domain adaptation (DA) setting named Open-Set Multi-Source Multi-Target Domain Adaptation (OS-nSmT), with n and m being number of source and target domains respectively. Next, we propose a graph attention based framework named DEGAA which can capture information from multiple source and target domains without knowing the exact label-set of the target. We argue that our method, though offered for multiple sources and multiple targets, can also be agnostic to various other DA settings. To check the robustness and versatility of DEGAA, we put forward ample experiments and ablation studies.
comment: Submitted in NeurIPS 2021 Workshop on Pre-registration in Machine Learning
☆ Hand Pose Estimation via Multiview Collaborative Self-Supervised Learning
3D hand pose estimation has made significant progress in recent years. However, the improvement is highly dependent on the emergence of large-scale annotated datasets. To alleviate the label-hungry limitation, we propose a multi-view collaborative self-supervised learning framework, HaMuCo, that estimates hand pose only with pseudo labels for training. We use a two-stage strategy to tackle the noisy label challenge and the multi-view ``groupthink'' problem. In the first stage, we estimate the 3D hand poses for each view independently. In the second stage, we employ a cross-view interaction network to capture the cross-view correlated features and use multi-view consistency loss to achieve collaborative learning among views. To further enhance the collaboration between single-view and multi-view, we fuse the results of all views to supervise the single-view network. To summarize, we introduce collaborative learning in two folds, the cross-view level and the multi- to single-view level. Extensive experiments show that our method can achieve state-of-the-art performance on multi-view self-supervised hand pose estimation. Moreover, ablation studies verify the effectiveness of each component. Results on multiple datasets further demonstrate the generalization ability of our network.
☆ Domain Generalization Emerges from Dreaming
Recent studies have proven that DNNs, unlike human vision, tend to exploit texture information rather than shape. Such texture bias is one of the factors for the poor generalization performance of DNNs. We observe that the texture bias negatively affects not only in-domain generalization but also out-of-distribution generalization, i.e., Domain Generalization. Motivated by the observation, we propose a new framework to reduce the texture bias of a model by a novel optimization-based data augmentation, dubbed Stylized Dream. Our framework utilizes adaptive instance normalization (AdaIN) to augment the style of an original image yet preserve the content. We then adopt a regularization loss to predict consistent outputs between Stylized Dream and original images, which encourages the model to learn shape-based representations. Extensive experiments show that the proposed method achieves state-of-the-art performance in out-of-distribution settings on public benchmark datasets: PACS, VLCS, OfficeHome, TerraIncognita, and DomainNet.
comment: 23 pages, 4 figures
☆ Deep-Learning Tool for Early Identifying Non-Traumatic Intracranial Hemorrhage Etiology based on CT Scan
Background: To develop an artificial intelligence system that can accurately identify acute non-traumatic intracranial hemorrhage (ICH) etiology based on non-contrast CT (NCCT) scans and investigate whether clinicians can benefit from it in a diagnostic setting. Materials and Methods: The deep learning model was developed with 1868 eligible NCCT scans with non-traumatic ICH collected between January 2011 and April 2018. We tested the model on two independent datasets (TT200 and SD 98) collected after April 2018. The model's diagnostic performance was compared with clinicians's performance. We further designed a simulated study to compare the clinicians's performance with and without the deep learning system augmentation. Results: The proposed deep learning system achieved area under the receiver operating curve of 0.986 (95% CI 0.967-1.000) on aneurysms, 0.952 (0.917-0.987) on hypertensive hemorrhage, 0.950 (0.860-1.000) on arteriovenous malformation (AVM), 0.749 (0.586-0.912) on Moyamoya disease (MMD), 0.837 (0.704-0.969) on cavernous malformation (CM), and 0.839 (0.722-0.959) on other causes in TT200 dataset. Given a 90% specificity level, the sensitivities of our model were 97.1% and 90.9% for aneurysm and AVM diagnosis, respectively. The model also shows an impressive generalizability in an independent dataset SD98. The clinicians achieve significant improvements in the sensitivity, specificity, and accuracy of diagnoses of certain hemorrhage etiologies with proposed system augmentation. Conclusions: The proposed deep learning algorithms can be an effective tool for early identification of hemorrhage etiologies based on NCCT scans. It may also provide more information for clinicians for triage and further imaging examination selection.
☆ QR-CLIP: Introducing Explicit Open-World Knowledge for Location and Time Reasoning
Daily images may convey abstract meanings that require us to memorize and infer profound information from them. To encourage such human-like reasoning, in this work, we teach machines to predict where and when it was taken rather than performing basic tasks like traditional segmentation or classification. Inspired by Horn's QR theory, we designed a novel QR-CLIP model consisting of two components: 1) the Quantity module first retrospects more open-world knowledge as the candidate language inputs; 2) the Relevance module carefully estimates vision and language cues and infers the location and time. Experiments show our QR-CLIP's effectiveness, and it outperforms the previous SOTA on each task by an average of about 10% and 130% relative lift in terms of location and time reasoning. This study lays a technical foundation for location and time reasoning and suggests that effectively introducing open-world knowledge is one of the panaceas for the tasks.
comment: On-Processing Work
☆ An optimization method for out-of-distribution anomaly detection models
Frequent false alarms impede the promotion of unsupervised anomaly detection algorithms in industrial applications. Potential characteristics of false alarms depending on the trained detector are revealed by investigating density probability distributions of prediction scores in the out-of-distribution anomaly detection tasks. An SVM-based classifier is exploited as a post-processing module to identify false alarms from the anomaly map at the object level. Besides, a sample synthesis strategy is devised to incorporate fuzzy prior knowledge on the specific application in the anomaly-free training dataset. Experimental results illustrate that the proposed method comprehensively improves the performances of two segmentation models at both image and pixel levels on two industrial applications.
☆ Adaptive Siamese Tracking with a Compact Latent Network ECCV 2020
In this paper, we provide an intuitive viewing to simplify the Siamese-based trackers by converting the tracking task to a classification. Under this viewing, we perform an in-depth analysis for them through visual simulations and real tracking examples, and find that the failure cases in some challenging situations can be regarded as the issue of missing decisive samples in offline training. Since the samples in the initial (first) frame contain rich sequence-specific information, we can regard them as the decisive samples to represent the whole sequence. To quickly adapt the base model to new scenes, a compact latent network is presented via fully using these decisive samples. Specifically, we present a statistics-based compact latent feature for fast adjustment by efficiently extracting the sequence-specific information. Furthermore, a new diverse sample mining strategy is designed for training to further improve the discrimination ability of the proposed compact latent network. Finally, a conditional updating strategy is proposed to efficiently update the basic models to handle scene variation during the tracking phase. To evaluate the generalization ability and effectiveness and of our method, we apply it to adjust three classical Siamese-based trackers, namely SiamRPN++, SiamFC, and SiamBAN. Extensive experimental results on six recent datasets demonstrate that all three adjusted trackers obtain the superior performance in terms of the accuracy, while having high running speed.
comment: Extended version of the paper "CLNet: A Compact Latent Network for Fast Adjusting Siamese Trackers" presented at ECCV 2020. Accepted at TPAMI
☆ Multimodal Chain-of-Thought Reasoning in Language Models
Large language models (LLMs) have shown impressive performance on complex reasoning by leveraging chain-of-thought (CoT) prompting to generate intermediate reasoning chains as the rationale to infer the answer. However, existing CoT studies are mostly isolated in the language modality with LLMs, where LLMs are hard to deploy. To elicit CoT reasoning in multimodality, a possible solution is to fine-tune small language models by fusing the vision and language features to perform CoT reasoning. The key challenge is that those language models tend to generate hallucinated reasoning chains that mislead the answer inference. To mitigate the effect of such mistakes, we propose Multimodal-CoT that incorporates vision features in a decoupled training framework. The framework separates the rationale generation and answer inference into two stages. By incorporating the vision features in both stages, the model is able to generate effective rationales that contribute to answer inference. With Multimodal-CoT, our model under 1 billion parameters outperforms the previous state-of-the-art LLM (GPT-3.5) by 16% (75.17%->91.68%) on the ScienceQA benchmark and even surpasses human performance. Code is publicly available at https://github.com/amazon-science/mm-cot.
☆ Visual Realism Assessment for Face-swap Videos
Deep-learning based face-swap videos, also known as deep fakes, are becoming more and more realistic and deceiving. The malicious usage of these face-swap videos has caused wide concerns. The research community has been focusing on the automatic detection of these fake videos, but the as sessment of their visual realism, as perceived by human eyes, is still an unexplored dimension. Visual realism assessment, or VRA, is essential for assessing the potential impact that may be brought by a specific face-swap video, and it is also important as a quality assessment metric to compare different face-swap methods. In this paper, we make a small step to wards this new VRA direction by building a benchmark for evaluating the effectiveness of different automatic VRA models, which range from using traditional hand-crafted features to different kinds of deep-learning features. The evaluations are based on a recent competition dataset named as DFGC 2022, which contains 1400 diverse face-swap videos that are annotated with Mean Opinion Scores (MOS) on visual realism. Comprehensive experiment results using 11 models and 3 protocols are shown and discussed. We demonstrate the feasibility of devising effective VRA models for assessing face-swap videos and methods. The particular usefulness of existing deepfake detection features for VRA is also noted. The code and benchmark will be made publicly available.
☆ Cooperative Saliency-based Obstacle Detection and AR Rendering for Increased Situational Awareness
Autonomous vehicles are expected to operate safely in real-life road conditions in the next years. Nevertheless, unanticipated events such as the existence of unexpected objects in the range of the road, can put safety at risk. The advancement of sensing and communication technologies and Internet of Things may facilitate the recognition of hazardous situations and information exchange in a cooperative driving scheme, providing new opportunities for the increase of collaborative situational awareness. Safe and unobtrusive visualization of the obtained information may nowadays be enabled through the adoption of novel Augmented Reality (AR) interfaces in the form of windshields. Motivated by these technological opportunities, we propose in this work a saliency-based distributed, cooperative obstacle detection and rendering scheme for increasing the driver's situational awareness through (i) automated obstacle detection, (ii) AR visualization and (iii) information sharing (upcoming potential dangers) with other connected vehicles or road infrastructure. An extensive evaluation study using a variety of real datasets for pothole detection showed that the proposed method provides favorable results and features compared to other recent and relevant approaches.
☆ Advances and Challenges in Multimodal Remote Sensing Image Registration
Over the past few decades, with the rapid development of global aerospace and aerial remote sensing technology, the types of sensors have evolved from the traditional monomodal sensors (e.g., optical sensors) to the new generation of multimodal sensors [e.g., multispectral, hyperspectral, light detection and ranging (LiDAR) and synthetic aperture radar (SAR) sensors]. These advanced devices can dynamically provide various and abundant multimodal remote sensing images with different spatial, temporal, and spectral resolutions according to different application requirements. Since then, it is of great scientific significance to carry out the research of multimodal remote sensing image registration, which is a crucial step for integrating the complementary information among multimodal data and making comprehensive observations and analysis of the Earths surface. In this work, we will present our own contributions to the field of multimodal image registration, summarize the advantages and limitations of existing multimodal image registration methods, and then discuss the remaining challenges and make a forward-looking prospect for the future development of the field.
comment: 10 pages, 4 figures
☆ GANalyzer: Analysis and Manipulation of GANs Latent Space for Controllable Face Synthesis
Generative Adversarial Networks (GANs) are capable of synthesizing high-quality facial images. Despite their success, GANs do not provide any information about the relationship between the input vectors and the generated images. Currently, facial GANs are trained on imbalanced datasets, which generate less diverse images. For example, more than 77% of 100K images that we randomly synthesized using the StyleGAN3 are classified as Happy, and only around 3% are Angry. The problem even becomes worse when a mixture of facial attributes is desired: less than 1% of the generated samples are Angry Woman, and only around 2% are Happy Black. To address these problems, this paper proposes a framework, called GANalyzer, for the analysis, and manipulation of the latent space of well-trained GANs. GANalyzer consists of a set of transformation functions designed to manipulate latent vectors for a specific facial attribute such as facial Expression, Age, Gender, and Race. We analyze facial attribute entanglement in the latent space of GANs and apply the proposed transformation for editing the disentangled facial attributes. Our experimental results demonstrate the strength of GANalyzer in editing facial attributes and generating any desired faces. We also create and release a balanced photo-realistic human face dataset. Our code is publicly available on GitHub.
☆ No One Left Behind: Real-World Federated Class-Incremental Learning
Federated learning (FL) is a hot collaborative training framework via aggregating model parameters of decentralized local clients. However, most existing models unreasonably assume that data categories of FL framework are known and fxed in advance. It renders the global model to signifcantly degrade recognition performance on old categories (i.e., catastrophic forgetting), when local clients receive new categories consecutively under limited memory of storing old categories. Moreover, some new local clients that collect novel categories unseen by other clients may be introduced to the FL training irregularly, which further exacerbates the catastrophic forgetting on old categories. To tackle the above issues, we propose a novel Local-Global Anti-forgetting (LGA) model to address local and global catastrophic forgetting on old categories, which is a pioneering work to explore a global class-incremental model in the FL feld. Specifcally, considering tackling class imbalance of local client to surmount local forgetting, we develop a category-balanced gradient-adaptive compensation loss and a category gradient-induced semantic distillation loss. They can balance heterogeneous forgetting speeds of hard-to-forget and easy-to-forget old categories, while ensure intrinsic class relations consistency within different incremental tasks. Moreover, a proxy server is designed to tackle global forgetting caused by Non-IID class imbalance between different clients. It collects perturbed prototype images of new categories from local clients via prototype gradient communication under privacy preservation, and augments them via self-supervised prototype augmentation to choose the best old global model and improve local distillation gain. Experiments on representative datasets verify superior performance of our model against other comparison methods.
comment: 16 pages, 9 figures
☆ Language Quantized AutoEncoders: Towards Unsupervised Text-Image Alignment
Recent progress in scaling up large language models has shown impressive capabilities in performing few-shot learning across a wide range of text-based tasks. However, a key limitation is that these language models fundamentally lack visual perception - a crucial attribute needed to extend these models to be able to interact with the real world and solve vision tasks, such as in visual-question answering and robotics. Prior works have largely connected image to text through pretraining and/or fine-tuning on curated image-text datasets, which can be a costly and expensive process. In order to resolve this limitation, we propose a simple yet effective approach called Language-Quantized AutoEncoder (LQAE), a modification of VQ-VAE that learns to align text-image data in an unsupervised manner by leveraging pretrained language models (e.g., BERT, RoBERTa). Our main idea is to encode image as sequences of text tokens by directly quantizing image embeddings using a pretrained language codebook. We then apply random masking followed by a BERT model, and have the decoder reconstruct the original image from BERT predicted text token embeddings. By doing so, LQAE learns to represent similar images with similar clusters of text tokens, thereby aligning these two modalities without the use of aligned text-image pairs. This enables few-shot image classification with large language models (e.g., GPT-3) as well as linear classification of images based on BERT text features. To the best of our knowledge, our work is the first work that uses unaligned images for multimodal tasks by leveraging the power of pretrained language models.
comment: arXiv admin note: text overlap with arXiv:2106.13884 by other authors
☆ Longformer: Longitudinal Transformer for Alzheimer's Disease Classification with Structural MRIs
Structural magnetic resonance imaging (sMRI) is widely used for brain neurological disease diagnosis; while longitudinal MRIs are often collected to monitor and capture disease progression, as clinically used in diagnosing Alzheimer's disease (AD). However, most current methods neglect AD's progressive nature and only take a single sMRI for recognizing AD. In this paper, we consider the problem of leveraging the longitudinal MRIs of a subject for AD identification. To capture longitudinal changes in sMRIs, we propose a novel model Longformer, a spatiotemporal transformer network that performs attention mechanisms spatially on sMRIs at each time point and integrates brain region features over time to obtain longitudinal embeddings for classification. Our Longformer achieves state-of-the-art performance on two binary classification tasks of separating different stages of AD using the ADNI dataset. Our source code is available at https://github.com/Qybc/LongFormer.
comment: 9 pages, 5 figures
☆ KST-Mixer: Kinematic Spatio-Temporal Data Mixer For Colon Shape Estimation MICCAI
We propose a spatio-temporal mixing kinematic data estimation method to estimate the shape of the colon with deformations caused by colonoscope insertion. Endoscope tracking or a navigation system that navigates physicians to target positions is needed to reduce such complications as organ perforations. Although many previous methods focused to track bronchoscopes and surgical endoscopes, few number of colonoscope tracking methods were proposed. This is because the colon largely deforms during colonoscope insertion. The deformation causes significant tracking errors. Colon deformation should be taken into account in the tracking process. We propose a colon shape estimation method using a Kinematic Spatio-Temporal data Mixer (KST-Mixer) that can be used during colonoscope insertions to the colon. Kinematic data of a colonoscope and the colon, including positions and directions of their centerlines, are obtained using electromagnetic and depth sensors. The proposed method separates the data into sub-groups along the spatial and temporal axes. The KST-Mixer extracts kinematic features and mix them along the spatial and temporal axes multiple times. We evaluated colon shape estimation accuracies in phantom studies. The proposed method achieved 11.92 mm mean Euclidean distance error, the smallest of the previous methods. Statistical analysis indicated that the proposed method significantly reduced the error compared to the previous methods.
comment: Accepted paper as an oral presentation at Joint MICCAI workshop 2022, AE-CAI/CARE/OR2.0. Received the Outstanding Paper Award
☆ AOP-Net: All-in-One Perception Network for Joint LiDAR-based 3D Object Detection and Panoptic Segmentation
LiDAR-based 3D object detection and panoptic segmentation are two crucial tasks in the perception systems of autonomous vehicles and robots. In this paper, we propose All-in-One Perception Network (AOP-Net), a LiDAR-based multi-task framework that combines 3D object detection and panoptic segmentation. In this method, a dual-task 3D backbone is developed to extract both panoptic- and detection-level features from the input LiDAR point cloud. Also, a new 2D backbone that intertwines Multi-Layer Perceptron (MLP) and convolution layers is designed to further improve the detection task performance. Finally, a novel module is proposed to guide the detection head by recovering useful features discarded during down-sampling operations in the 3D backbone. This module leverages estimated instance segmentation masks to recover detailed information from each candidate object. The AOP-Net achieves state-of-the-art performance for published works on the nuScenes benchmark for both 3D object detection and panoptic segmentation tasks. Also, experiments show that our method easily adapts to and significantly improves the performance of any BEV-based 3D object detection method.
comment: Under review
☆ Exploring Invariant Representation for Visible-Infrared Person Re-Identification
Cross-spectral person re-identification, which aims to associate identities to pedestrians across different spectra, faces a main challenge of the modality discrepancy. In this paper, we address the problem from both image-level and feature-level in an end-to-end hybrid learning framework named robust feature mining network (RFM). In particular, we observe that the reflective intensity of the same surface in photos shot in different wavelengths could be transformed using a linear model. Besides, we show the variable linear factor across the different surfaces is the main culprit which initiates the modality discrepancy. We integrate such a reflection observation into an image-level data augmentation by proposing the linear transformation generator (LTG). Moreover, at the feature level, we introduce a cross-center loss to explore a more compact intra-class distribution and modality-aware spatial attention to take advantage of textured regions more efficiently. Experiment results on two standard cross-spectral person re-identification datasets, i.e., RegDB and SYSU-MM01, have demonstrated state-of-the-art performance.
☆ Vision Transformer-based Feature Extraction for Generalized Zero-Shot Learning
Generalized zero-shot learning (GZSL) is a technique to train a deep learning model to identify unseen classes using the image attribute. In this paper, we put forth a new GZSL approach exploiting Vision Transformer (ViT) to maximize the attribute-related information contained in the image feature. In ViT, the entire image region is processed without the degradation of the image resolution and the local image information is preserved in patch features. To fully enjoy these benefits of ViT, we exploit patch features as well as the CLS feature in extracting the attribute-related image feature. In particular, we propose a novel attention-based module, called attribute attention module (AAM), to aggregate the attribute-related information in patch features. In AAM, the correlation between each patch feature and the synthetic image attribute is used as the importance weight for each patch. From extensive experiments on benchmark datasets, we demonstrate that the proposed technique outperforms the state-of-the-art GZSL approaches by a large margin.
comment: 21 pages, 10 figures
☆ Disentanglement of Latent Representations via Sparse Causal Interventions IJCAI 2023
The process of generating data such as images is controlled by independent and unknown factors of variation. The retrieval of these variables has been studied extensively in the disentanglement, causal representation learning, and independent component analysis fields. Recently, approaches merging these domains together have shown great success. Instead of directly representing the factors of variation, the problem of disentanglement can be seen as finding the interventions on one image that yield a change to a single factor. Following this assumption, we introduce a new method for disentanglement inspired by causal dynamics that combines causality theory with vector-quantized variational autoencoders. Our model considers the quantized vectors as causal variables and links them in a causal graph. It performs causal interventions on the graph and generates atomic transitions affecting a unique factor of variation in the image. We also introduce a new task of action retrieval that consists of finding the action responsible for the transition between two images. We test our method on standard synthetic and real-world disentanglement datasets. We show that it can effectively disentangle the factors of variation and perform precise interventions on high-level semantic attributes of an image without affecting its quality, even with imbalanced data distributions.
comment: 16 pages, 10 pages for the main paper and 6 pages for the supplement, 14 figures, submitted to IJCAI 2023
☆ CLIPood: Generalizing CLIP to Out-of-Distributions
Out-of-distribution (OOD) generalization, where the model needs to handle distribution shifts from training, is a major challenge of machine learning. Recently, contrastive language-image pre-training (CLIP) models have shown impressive zero-shot ability, revealing a promising path toward OOD generalization. However, to boost upon zero-shot performance, further adaptation of CLIP on downstream tasks is indispensable but undesirably degrades OOD generalization ability. In this paper, we aim at generalizing CLIP to out-of-distribution test data on downstream tasks. Beyond the two canonical OOD situations, domain shift and open class, we tackle a more general but difficult in-the-wild setting where both OOD situations may occur on the unseen test data. We propose CLIPood, a simple fine-tuning method that can adapt CLIP models to all OOD situations. To exploit semantic relations between classes from the text modality, CLIPood introduces a new training objective, margin metric softmax (MMS), with class adaptive margins for fine-tuning. Moreover, to incorporate both the pre-trained zero-shot model and the fine-tuned task-adaptive model, CLIPood proposes a new Beta moving average (BMA) to maintain a temporal ensemble according to Beta distribution. Experiments on diverse datasets with different OOD scenarios show that CLIPood consistently outperforms existing generalization techniques.
☆ Online Continual Learning via the Knowledge Invariant and Spread-out Properties
The goal of continual learning is to provide intelligent agents that are capable of learning continually a sequence of tasks using the knowledge obtained from previous tasks while performing well on prior tasks. However, a key challenge in this continual learning paradigm is catastrophic forgetting, namely adapting a model to new tasks often leads to severe performance degradation on prior tasks. Current memory-based approaches show their success in alleviating the catastrophic forgetting problem by replaying examples from past tasks when new tasks are learned. However, these methods are infeasible to transfer the structural knowledge from previous tasks i.e., similarities or dissimilarities between different instances. Furthermore, the learning bias between the current and prior tasks is also an urgent problem that should be solved. In this work, we propose a new method, named Online Continual Learning via the Knowledge Invariant and Spread-out Properties (OCLKISP), in which we constrain the evolution of the embedding features via Knowledge Invariant and Spread-out Properties (KISP). Thus, we can further transfer the inter-instance structural knowledge of previous tasks while alleviating the forgetting due to the learning bias. We empirically evaluate our proposed method on four popular benchmarks for continual learning: Split CIFAR 100, Split SVHN, Split CUB200 and Split Tiny-Image-Net. The experimental results show the efficacy of our proposed method compared to the state-of-the-art continual learning algorithms.
comment: 30 pages
☆ SHINE: Deep Learning-Based Accessible Parking Management System
The enhancement of science and technology has helped expand urban cities like never before. Due to the undeniable benefits of owning a private vehicle, the number of cars has rocketed in many parts of the world, including South Korea. However, these gradual increments in the number of vehicles lead to parking-related problems, including the abuse of disabled parking spaces (referred to as accessible parking spaces hereafter). Due to the high frame rate of surveillance cameras, traditional license plate recognition (LPR) systems are ineffective in real-time. On the other hand, natural and artificial noise and differences in lighting and weather conditions make detection and recognition difficult for these systems. With the growing concept of parking 4.0, many sensors, IoT and deep learning-based approaches have been applied to automatic LPR and parking management systems. However, the studies show a need for a robust and efficient model for managing accessible parking spaces in South Korea. We have proposed a novel system called 'SHINE', which uses the deep learning-based object detection algorithm for detecting the vehicle, license plate, and disability badges (referred to as cards, badges, or access badges hereafter) and then authenticates the rights to use the accessible parking spaces by coordinating with the central server. This model, achieving 92.16% mean average precision, is believed to solve the problem of accessible parking space abuse.
comment: 14 papges main manuscript, 6 pages Appendix, 8 figures
☆ RobustNeRF: Ignoring Distractors with Robust Losses
Neural radiance fields (NeRF) excel at synthesizing new views given multi-view, calibrated images of a static scene. When scenes include distractors, which are not persistent during image capture (moving objects, lighting variations, shadows), artifacts appear as view-dependent effects or 'floaters'. To cope with distractors, we advocate a form of robust estimation for NeRF training, modeling distractors in training data as outliers of an optimization problem. Our method successfully removes outliers from a scene and improves upon our baselines, on synthetic and real-world scenes. Our technique is simple to incorporate in modern NeRF frameworks, with few hyper-parameters. It does not assume a priori knowledge of the types of distractors, and is instead focused on the optimization problem rather than pre-processing or modeling transient objects. More results on our page https://robustnerf.github.io/public.
☆ SpaceYOLO: A Human-Inspired Model for Real-time, On-board Spacecraft Feature Detection
The rapid proliferation of non-cooperative spacecraft and space debris in orbit has precipitated a surging demand for on-orbit servicing and space debris removal at a scale that only autonomous missions can address, but the prerequisite autonomous navigation and flightpath planning to safely capture an unknown, non-cooperative, tumbling space object is an open problem. This requires algorithms for real-time, automated spacecraft feature recognition to pinpoint the locations of collision hazards (e.g. solar panels or antennas) and safe docking features (e.g. satellite bodies or thrusters) so safe, effective flightpaths can be planned. Prior work in this area reveals the performance of computer vision models are highly dependent on the training dataset and its coverage of scenarios visually similar to the real scenarios that occur in deployment. Hence, the algorithm may have degraded performance under certain lighting conditions even when the rendezvous maneuver conditions of the chaser to the target spacecraft are the same. This work delves into how humans perform these tasks through a survey of how aerospace engineering students experienced with spacecraft shapes and components recognize features of the three spacecraft: Landsat, Envisat, Anik, and the orbiter Mir. The survey reveals that the most common patterns in the human detection process were to consider the shape and texture of the features: antennas, solar panels, thrusters, and satellite bodies. This work introduces a novel algorithm SpaceYOLO, which fuses a state-of-the-art object detector YOLOv5 with a separate neural network based on these human-inspired decision processes exploiting shape and texture. Performance in autonomous spacecraft detection of SpaceYOLO is compared to ordinary YOLOv5 in hardware-in-the-loop experiments under different lighting and chaser maneuver conditions at the ORION Laboratory at Florida Tech.
comment: Accepted at IEEE Aerospace Conference 2023, 11 pages, 21 figures
☆ Dynamic Atomic Column Detection in Transmission Electron Microscopy Videos via Ridge Estimation
Ridge detection is a classical tool to extract curvilinear features in image processing. As such, it has great promise in applications to material science problems; specifically, for trend filtering relatively stable atom-shaped objects in image sequences, such as Transmission Electron Microscopy (TEM) videos. Standard analysis of TEM videos is limited to frame-by-frame object recognition. We instead harness temporal correlation across frames through simultaneous analysis of long image sequences, specified as a spatio-temporal image tensor. We define new ridge detection algorithms to non-parametrically estimate explicit trajectories of atomic-level object locations as a continuous function of time. Our approach is specially tailored to handle temporal analysis of objects that seemingly stochastically disappear and subsequently reappear throughout a sequence. We demonstrate that the proposed method is highly effective and efficient in simulation scenarios, and delivers notable performance improvements in TEM experiments compared to other material science benchmarks.
comment: 27 pages, 11 figures
☆ A sliced-Wasserstein distance-based approach for out-of-class-distribution detection
There exist growing interests in intelligent systems for numerous medical imaging, image processing, and computer vision applications, such as face recognition, medical diagnosis, character recognition, and self-driving cars, among others. These applications usually require solving complex classification problems involving complex images with unknown data generative processes. In addition to recent successes of the current classification approaches relying on feature engineering and deep learning, several shortcomings of them, such as the lack of robustness, generalizability, and interpretability, have also been observed. These methods often require extensive training data, are computationally expensive, and are vulnerable to out-of-distribution samples, e.g., adversarial attacks. Recently, an accurate, data-efficient, computationally efficient, and robust transport-based classification approach has been proposed, which describes a generative model-based problem formulation and closed-form solution for a specific category of classification problems. However, all these approaches lack mechanisms to detect test samples outside the class distributions used during training. In real-world settings, where the collected training samples are unable to exhaust or cover all classes, the traditional classification schemes are unable to handle the unseen classes effectively, which is especially an important issue for safety-critical systems, such as self-driving and medical imaging diagnosis. In this work, we propose a method for detecting out-of-class distributions based on the distribution of sliced-Wasserstein distance from the Radon Cumulative Distribution Transform (R-CDT) subspace. We tested our method on the MNIST and two medical image datasets and reported better accuracy than the state-of-the-art methods without an out-of-class distribution detection procedure.
☆ CTE: A Dataset for Contextualized Table Extraction
Relevant information in documents is often summarized in tables, helping the reader to identify useful facts. Most benchmark datasets support either document layout analysis or table understanding, but lack in providing data to apply both tasks in a unified way. We define the task of Contextualized Table Extraction (CTE), which aims to extract and define the structure of tables considering the textual context of the document. The dataset comprises 75k fully annotated pages of scientific papers, including more than 35k tables. Data are gathered from PubMed Central, merging the information provided by annotations in the PubTables-1M and PubLayNet datasets. The dataset can support CTE and adds new classes to the original ones. The generated annotations can be used to develop end-to-end pipelines for various tasks, including document layout analysis, table detection, structure recognition, and functional analysis. We formally define CTE and evaluation metrics, showing which subtasks can be tackled, describing advantages, limitations, and future works of this collection of data. Annotations and code will be accessible a https://github.com/AILab-UniFI/cte-dataset.
☆ Benchmarking Probabilistic Deep Learning Methods for License Plate Recognition
Learning-based algorithms for automated license plate recognition implicitly assume that the training and test data are well aligned. However, this may not be the case under extreme environmental conditions, or in forensic applications where the system cannot be trained for a specific acquisition device. Predictions on such out-of-distribution images have an increased chance of failing. But this failure case is oftentimes hard to recognize for a human operator or an automated system. Hence, in this work we propose to model the prediction uncertainty for license plate recognition explicitly. Such an uncertainty measure allows to detect false predictions, indicating an analyst when not to trust the result of the automated license plate recognition. In this paper, we compare three methods for uncertainty quantification on two architectures. The experiments on synthetic noisy or blurred low-resolution images show that the predictive uncertainty reliably finds wrong predictions. We also show that a multi-task combination of classification and super-resolution improves the recognition performance by 109\% and the detection of wrong predictions by 29 %.
☆ A Convolutional-based Model for Early Prediction of Alzheimer's based on the Dementia Stage in the MRI Brain Images
Alzheimer's disease is a degenerative brain disease. Being the primary cause of Dementia in adults and progressively destroys brain memory. Though Alzheimer's disease does not have a cure currently, diagnosing it at an earlier stage will help reduce the severity of the disease. Thus, early diagnosis of Alzheimer's could help to reduce or stop the disease from progressing. In this paper, we proposed a deep convolutional neural network-based model for learning model using to determine the stage of Dementia in adults based on the Magnetic Resonance Imaging (MRI) images to detect the early onset of Alzheimer's.
comment: Short paper, Under Review in FLAIRS-36
☆ Hyperbolic Contrastive Learning
Learning good image representations that are beneficial to downstream tasks is a challenging task in computer vision. As such, a wide variety of self-supervised learning approaches have been proposed. Among them, contrastive learning has shown competitive performance on several benchmark datasets. The embeddings of contrastive learning are arranged on a hypersphere that results in using the inner (dot) product as a distance measurement in Euclidean space. However, the underlying structure of many scientific fields like social networks, brain imaging, and computer graphics data exhibit highly non-Euclidean latent geometry. We propose a novel contrastive learning framework to learn semantic relationships in the hyperbolic space. Hyperbolic space is a continuous version of trees that naturally owns the ability to model hierarchical structures and is thus beneficial for efficient contrastive representation learning. We also extend the proposed Hyperbolic Contrastive Learning (HCL) to the supervised domain and studied the adversarial robustness of HCL. The comprehensive experiments show that our proposed method achieves better results on self-supervised pretraining, supervised classification, and higher robust accuracy than baseline methods.
Self-Supervised Relation Alignment for Scene Graph Generation
The goal of scene graph generation is to predict a graph from an input image, where nodes correspond to identified and localized objects and edges to their corresponding interaction predicates. Existing methods are trained in a fully supervised manner and focus on message passing mechanisms, loss functions, and/or bias mitigation. In this work we introduce a simple-yet-effective self-supervised relational alignment regularization designed to improve the scene graph generation performance. The proposed alignment is general and can be combined with any existing scene graph generation framework, where it is trained alongside the original model's objective. The alignment is achieved through distillation, where an auxiliary relation prediction branch, that mirrors and shares parameters with the supervised counterpart, is designed. In the auxiliary branch, relational input features are partially masked prior to message passing and predicate prediction. The predictions for masked relations are then aligned with the supervised counterparts after the message passing. We illustrate the effectiveness of this self-supervised relational alignment in conjunction with two scene graph generation architectures, SGTR and Neural Motifs, and show that in both cases we achieve significantly improved performance.
☆ MoE-Fusion: Instance Embedded Mixture-of-Experts for Infrared and Visible Image Fusion
Infrared and visible image fusion can compensate for the incompleteness of single-modality imaging and provide a more comprehensive scene description based on cross-modal complementarity. Most works focus on learning the overall cross-modal features by high- and low-frequency constraints at the image level alone, ignoring the fact that cross-modal instance-level features often contain more valuable information. To fill this gap, we model cross-modal instance-level features by embedding instance information into a set of Mixture-of-Experts (MoEs) for the first time, prompting image fusion networks to specifically learn instance-level information. We propose a novel framework with instance embedded Mixture-of-Experts for infrared and visible image fusion, termed MoE-Fusion, which contains an instance embedded MoE group (IE-MoE), an MoE-Decoder, two encoders, and two auxiliary detection networks. By embedding the instance-level information learned in the auxiliary network, IE-MoE achieves specialized learning of cross-modal foreground and background features. MoE-Decoder can adaptively select suitable experts for cross-modal feature decoding and obtain fusion results dynamically. Extensive experiments show that our MoE-Fusion outperforms state-of-the-art methods in preserving contrast and texture details by learning instance-level information in cross-modal images.
☆ Continual Learning with Scaled Gradient Projection AAAI 2023
In neural networks, continual learning results in gradient interference among sequential tasks, leading to catastrophic forgetting of old tasks while learning new ones. This issue is addressed in recent methods by storing the important gradient spaces for old tasks and updating the model orthogonally during new tasks. However, such restrictive orthogonal gradient updates hamper the learning capability of the new tasks resulting in sub-optimal performance. To improve new learning while minimizing forgetting, in this paper we propose a Scaled Gradient Projection (SGP) method, where we combine the orthogonal gradient projections with scaled gradient steps along the important gradient spaces for the past tasks. The degree of gradient scaling along these spaces depends on the importance of the bases spanning them. We propose an efficient method for computing and accumulating importance of these bases using the singular value decomposition of the input representations for each task. We conduct extensive experiments ranging from continual image classification to reinforcement learning tasks and report better performance with less training overhead than the state-of-the-art approaches.
comment: Accepted at AAAI 2023
☆ Energy-Inspired Self-Supervised Pretraining for Vision Models
Motivated by the fact that forward and backward passes of a deep network naturally form symmetric mappings between input and output representations, we introduce a simple yet effective self-supervised vision model pretraining framework inspired by energy-based models (EBMs). In the proposed framework, we model energy estimation and data restoration as the forward and backward passes of a single network without any auxiliary components, e.g., an extra decoder. For the forward pass, we fit a network to an energy function that assigns low energy scores to samples that belong to an unlabeled dataset, and high energy otherwise. For the backward pass, we restore data from corrupted versions iteratively using gradient-based optimization along the direction of energy minimization. In this way, we naturally fold the encoder-decoder architecture widely used in masked image modeling into the forward and backward passes of a single vision model. Thus, our framework now accepts a wide range of pretext tasks with different data corruption methods, and permits models to be pretrained from masked image modeling, patch sorting, and image restoration, including super-resolution, denoising, and colorization. We support our findings with extensive experiments, and show the proposed method delivers comparable and even better performance with remarkably fewer epochs of training compared to the state-of-the-art self-supervised vision model pretraining methods. Our findings shed light on further exploring self-supervised vision model pretraining and pretext tasks beyond masked image modeling.
☆ Effective Robustness against Natural Distribution Shifts for Models with Different Training Data
``Effective robustness'' measures the extra out-of-distribution (OOD) robustness beyond what can be predicted from the in-distribution (ID) performance. Existing effective robustness evaluations typically use a single test set such as ImageNet to evaluate ID accuracy. This becomes problematic when evaluating models trained on different data distributions, e.g., comparing models trained on ImageNet vs. zero-shot language-image pre-trained models trained on LAION. In this paper, we propose a new effective robustness evaluation metric to compare the effective robustness of models trained on different data distributions. To do this we control for the accuracy on multiple ID test sets that cover the training distributions for all the evaluated models. Our new evaluation metric provides a better estimate of the effectiveness robustness and explains the surprising effective robustness gains of zero-shot CLIP-like models exhibited when considering only one ID dataset, while the gains diminish under our evaluation.
♻ ☆ Aerial Image Object Detection With Vision Transformer Detector (ViTDet)
The past few years have seen an increased interest in aerial image object detection due to its critical value to large-scale geo-scientific research like environmental studies, urban planning, and intelligence monitoring. However, the task is very challenging due to the birds-eye view perspective, complex backgrounds, large and various image sizes, different appearances of objects, and the scarcity of well-annotated datasets. Recent advances in computer vision have shown promise tackling the challenge. Specifically, Vision Transformer Detector (ViTDet) was proposed to extract multi-scale features for object detection. The empirical study shows that ViTDet's simple design achieves good performance on natural scene images and can be easily embedded into any detector architecture. To date, ViTDet's potential benefit to challenging aerial image object detection has not been explored. Therefore, in our study, 25 experiments were carried out to evaluate the effectiveness of ViTDet for aerial image object detection on three well-known datasets: Airbus Aircraft, RarePlanes, and Dataset of Object DeTection in Aerial images (DOTA). Our results show that ViTDet can consistently outperform its convolutional neural network counterparts on horizontal bounding box (HBB) object detection by a large margin (up to 17% on average precision) and that it achieves the competitive performance for oriented bounding box (OBB) object detection. Our results also establish a baseline for future research.
♻ ☆ A Closer Look at Few-shot Classification Again
Few-shot classification consists of a training phase where a model is learned on a relatively large dataset and an adaptation phase where the learned model is adapted to previously-unseen tasks with limited labeled samples. In this paper, we empirically prove that the training algorithm and the adaptation algorithm can be completely disentangled, which allows algorithm analysis and design to be done individually for each phase. Our meta-analysis for each phase reveals several interesting insights that may help better understand key aspects of few-shot classification and connections with other fields such as visual representation learning and transfer learning. We hope the insights and research challenges revealed in this paper can inspire future work in related directions.
♻ ☆ UDepth: Fast Monocular Depth Estimation for Visually-guided Underwater Robots
In this paper, we present a fast monocular depth estimation method for enabling 3D perception capabilities of low-cost underwater robots. We formulate a novel end-to-end deep visual learning pipeline named UDepth, which incorporates domain knowledge of image formation characteristics of natural underwater scenes. First, we adapt a new input space from raw RGB image space by exploiting underwater light attenuation prior, and then devise a least-squared formulation for coarse pixel-wise depth prediction. Subsequently, we extend this into a domain projection loss that guides the end-to-end learning of UDepth on over 9K RGB-D training samples. UDepth is designed with a computationally light MobileNetV2 backbone and a Transformer-based optimizer for ensuring fast inference rates on embedded systems. By domain-aware design choices and through comprehensive experimental analyses, we demonstrate that it is possible to achieve state-of-the-art depth estimation performance while ensuring a small computational footprint. Specifically, with 70%-80% less network parameters than existing benchmarks, UDepth achieves comparable and often better depth estimation performance. While the full model offers over 66 FPS (13 FPS) inference rates on a single GPU (CPU core), our domain projection for coarse depth prediction runs at 51.5 FPS rates on single-board NVIDIA Jetson TX2s. The inference pipelines are available at https://github.com/uf-robopi/UDepth.
comment: 10 pages, 6 figures
♻ ☆ BridgeTower: Building Bridges Between Encoders in Vision-Language Representation Learning AAAI 2023
Vision-Language (VL) models with the Two-Tower architecture have dominated visual-language representation learning in recent years. Current VL models either use lightweight uni-modal encoders and learn to extract, align and fuse both modalities simultaneously in a deep cross-modal encoder, or feed the last-layer uni-modal representations from the deep pre-trained uni-modal encoders into the top cross-modal encoder. Both approaches potentially restrict vision-language representation learning and limit model performance. In this paper, we propose BridgeTower, which introduces multiple bridge layers that build a connection between the top layers of uni-modal encoders and each layer of the cross-modal encoder. This enables effective bottom-up cross-modal alignment and fusion between visual and textual representations of different semantic levels of pre-trained uni-modal encoders in the cross-modal encoder. Pre-trained with only 4M images, BridgeTower achieves state-of-the-art performance on various downstream vision-language tasks. In particular, on the VQAv2 test-std set, BridgeTower achieves an accuracy of 78.73%, outperforming the previous state-of-the-art model METER by 1.09% with the same pre-training data and almost negligible additional parameters and computational costs. Notably, when further scaling the model, BridgeTower achieves an accuracy of 81.15%, surpassing models that are pre-trained on orders-of-magnitude larger datasets. Code and checkpoints are available at https://github.com/microsoft/BridgeTower.
comment: Accepted by AAAI 2023, Oral
♻ ☆ A Light Recipe to Train Robust Vision Transformers
In this paper, we ask whether Vision Transformers (ViTs) can serve as an underlying architecture for improving the adversarial robustness of machine learning models against evasion attacks. While earlier works have focused on improving Convolutional Neural Networks, we show that also ViTs are highly suitable for adversarial training to achieve competitive performance. We achieve this objective using a custom adversarial training recipe, discovered using rigorous ablation studies on a subset of the ImageNet dataset. The canonical training recipe for ViTs recommends strong data augmentation, in part to compensate for the lack of vision inductive bias of attention modules, when compared to convolutions. We show that this recipe achieves suboptimal performance when used for adversarial training. In contrast, we find that omitting all heavy data augmentation, and adding some additional bag-of-tricks ($\varepsilon$-warmup and larger weight decay), significantly boosts the performance of robust ViTs. We show that our recipe generalizes to different classes of ViT architectures and large-scale models on full ImageNet-1k. Additionally, investigating the reasons for the robustness of our models, we show that it is easier to generate strong attacks during training when using our recipe and that this leads to better robustness at test time. Finally, we further study one consequence of adversarial training by proposing a way to quantify the semantic nature of adversarial perturbations and highlight its correlation with the robustness of the model. Overall, we recommend that the community should avoid translating the canonical training recipes in ViTs to robust training and rethink common training choices in the context of adversarial training.
comment: Camera-ready version for SaTML 2023, code available at https://github.com/dedeswim/vits-robustness-torch
♻ ☆ Planckian Jitter: countering the color-crippling effects of color jitter on self-supervised training ICLR 2023
Several recent works on self-supervised learning are trained by mapping different augmentations of the same image to the same feature representation. The data augmentations used are of crucial importance to the quality of learned feature representations. In this paper, we analyze how the color jitter traditionally used in data augmentation negatively impacts the quality of the color features in learned feature representations. To address this problem, we propose a more realistic, physics-based color data augmentation - which we call Planckian Jitter - that creates realistic variations in chromaticity and produces a model robust to illumination changes that can be commonly observed in real life, while maintaining the ability to discriminate image content based on color information. Experiments confirm that such a representation is complementary to the representations learned with the currently-used color jitter augmentation and that a simple concatenation leads to significant performance gains on a wide range of downstream datasets. In addition, we present a color sensitivity analysis that documents the impact of different training methods on model neurons and shows that the performance of the learned features is robust with respect to illuminant variations.
comment: Accepted at Eleventh International Conference on Learning Representations (ICLR 2023)
♻ ☆ DiffMotion: Speech-Driven Gesture Synthesis Using Denoising Diffusion Model
Speech-driven gesture synthesis is a field of growing interest in virtual human creation. However, a critical challenge is the inherent intricate one-to-many mapping between speech and gestures. Previous studies have explored and achieved significant progress with generative models. Notwithstanding, most synthetic gestures are still vastly less natural. This paper presents DiffMotion, a novel speech-driven gesture synthesis architecture based on diffusion models. The model comprises an autoregressive temporal encoder and a denoising diffusion probability Module. The encoder extracts the temporal context of the speech input and historical gestures. The diffusion module learns a parameterized Markov chain to gradually convert a simple distribution into a complex distribution and generates the gestures according to the accompanied speech. Compared with baselines, objective and subjective evaluations confirm that our approach can produce natural and diverse gesticulation and demonstrate the benefits of diffusion-based models on speech-driven gesture synthesis.
comment: 13 pages, 3 figures
♻ ☆ Augmentation Component Analysis: Modeling Similarity via the Augmentation Overlaps ICLR 2023
Self-supervised learning aims to learn a embedding space where semantically similar samples are close. Contrastive learning methods pull views of samples together and push different samples away, which utilizes semantic invariance of augmentation but ignores the relationship between samples. To better exploit the power of augmentation, we observe that semantically similar samples are more likely to have similar augmented views. Therefore, we can take the augmented views as a special description of a sample. In this paper, we model such a description as the augmentation distribution and we call it augmentation feature. The similarity in augmentation feature reflects how much the views of two samples overlap and is related to their semantical similarity. Without computational burdens to explicitly estimate values of the augmentation feature, we propose Augmentation Component Analysis (ACA) with a contrastive-like loss to learn principal components and an on-the-fly projection loss to embed data. ACA equals an efficient dimension reduction by PCA and extracts low-dimensional embeddings, theoretically preserving the similarity of augmentation distribution between samples. Empirical results show our method can achieve competitive results against various traditional contrastive learning methods on different benchmarks.
comment: Accept to ICLR 2023
♻ ☆ A survey, review, and future trends of skin lesion segmentation and classification
The Computer-aided Diagnosis or Detection (CAD) approach for skin lesion analysis is an emerging field of research that has the potential to alleviate the burden and cost of skin cancer screening. Researchers have recently indicated increasing interest in developing such CAD systems, with the intention of providing a user-friendly tool to dermatologists to reduce the challenges encountered or associated with manual inspection. This article aims to provide a comprehensive literature survey and review of a total of 594 publications (356 for skin lesion segmentation and 238 for skin lesion classification) published between 2011 and 2022. These articles are analyzed and summarized in a number of different ways to contribute vital information regarding the methods for the development of CAD systems. These ways include relevant and essential definitions and theories, input data (dataset utilization, preprocessing, augmentations, and fixing imbalance problems), method configuration (techniques, architectures, module frameworks, and losses), training tactics (hyperparameter settings), and evaluation criteria. We intend to investigate a variety of performance-enhancing approaches, including ensemble and post-processing. We also discuss these dimensions to reveal their current trends based on utilization frequencies. In addition, we highlight the primary difficulties associated with evaluating skin lesion segmentation and classification systems using minimal datasets, as well as the potential solutions to these difficulties. Findings, recommendations, and trends are disclosed to inform future research on developing an automated and robust CAD system for skin lesion analysis.
comment: This manuscript has been accepted to be published in Computers in Biology and Medicine and has a total of 106 pages (single column and double spacing), 13 figures, and 11 tables
♻ ☆ INSPIRE: Intensity and spatial information-based deformable image registration
We present INSPIRE, a top-performing general-purpose method for deformable image registration. INSPIRE brings distance measures which combine intensity and spatial information into an elastic B-splines-based transformation model and incorporates an inverse inconsistency penalization supporting symmetric registration performance. We introduce several theoretical and algorithmic solutions which provide high computational efficiency and thereby applicability of the proposed framework in a wide range of real scenarios. We show that INSPIRE delivers highly accurate, as well as stable and robust registration results. We evaluate the method on a 2D dataset created from retinal images, characterized by presence of networks of thin structures. Here INSPIRE exhibits excellent performance, substantially outperforming the widely used reference methods. {We also evaluate INSPIRE on the Fundus Image Registration Dataset (FIRE), which consists of 134 pairs of separately acquired retinal images. INSPIRE exhibits excellent performance on the FIRE dataset, substantially outperforming several domain-specific methods.} We also evaluate the method on four benchmark datasets of 3D magnetic resonance images of brains, for a total of 2088 pairwise registrations. A comparison with 17 other state-of-the-art methods reveals that INSPIRE provides the best overall performance. Code is available at http://github.com/MIDA-group/inspire
comment: 24 pages, 7 figures, 5 tables
♻ ☆ Encouraging Intra-Class Diversity Through a Reverse Contrastive Loss for Better Single-Source Domain Generalization
Traditional deep learning algorithms often fail to generalize when they are tested outside of the domain of the training data. The issue can be mitigated by using unlabeled data from the target domain at training time, but because data distributions can change dynamically in real-life applications once a learned model is deployed, it is critical to create networks robust to unknown and unforeseen domain shifts. In this paper we focus on one of the reasons behind the inability of neural networks to be so: deep networks focus only on the most obvious, potentially spurious, clues to make their predictions and are blind to useful but slightly less efficient or more complex patterns. This behaviour has been identified and several methods partially addressed the issue. To investigate their effectiveness and limits, we first design a publicly available MNIST-based benchmark to precisely measure the ability of an algorithm to find the ''hidden'' patterns. Then, we evaluate state-of-the-art algorithms through our benchmark and show that the issue is largely unsolved. Finally, we propose a partially reversed contrastive loss to encourage intra-class diversity and find less strongly correlated patterns, whose efficiency is demonstrated by our experiments.
♻ ☆ Leveraging Modality-specific Representations for Audio-visual Speech Recognition via Reinforcement Learning AAAI2023
Audio-visual speech recognition (AVSR) has gained remarkable success for ameliorating the noise-robustness of speech recognition. Mainstream methods focus on fusing audio and visual inputs to obtain modality-invariant representations. However, such representations are prone to over-reliance on audio modality as it is much easier to recognize than video modality in clean conditions. As a result, the AVSR model underestimates the importance of visual stream in face of noise corruption. To this end, we leverage visual modality-specific representations to provide stable complementary information for the AVSR task. Specifically, we propose a reinforcement learning (RL) based framework called MSRL, where the agent dynamically harmonizes modality-invariant and modality-specific representations in the auto-regressive decoding process. We customize a reward function directly related to task-specific metrics (i.e., word error rate), which encourages the MSRL to effectively explore the optimal integration strategy. Experimental results on the LRS3 dataset show that the proposed method achieves state-of-the-art in both clean and various noisy conditions. Furthermore, we demonstrate the better generality of MSRL system than other baselines when test set contains unseen noises.
comment: Accepted by AAAI2023
♻ ☆ Information Prebuilt Recurrent Reconstruction Network for Video Super-Resolution
The video super-resolution (VSR) method based on the recurrent convolutional network has strong temporal modeling capability for video sequences. However, the temporal receptive field of different recurrent units in the unidirectional recurrent network is unbalanced. Earlier reconstruction frames receive less spatio-temporal information, resulting in fuzziness or artifacts. Although the bidirectional recurrent network can alleviate this problem, it requires more memory space and fails to perform many tasks with low latency requirements. To solve the above problems, we propose an end-to-end information prebuilt recurrent reconstruction network (IPRRN), consisting of an information prebuilt network (IPNet) and a recurrent reconstruction network (RRNet). By integrating sufficient information from the front of the video to build the hidden state needed for the initially recurrent unit to help restore the earlier frames, the information prebuilt network balances the input information difference at different time steps. In addition, we demonstrate an efficient recurrent reconstruction network, which outperforms the existing unidirectional recurrent schemes in all aspects. Many experiments have verified the effectiveness of the network we propose, which can effectively achieve better quantitative and qualitative evaluation performance compared to the existing state-of-the-art methods.
comment: 12 pages,9 figures. This work has been submitted to the IEEE for possible publication. Copyright may be transferred without notice, after which this version may no longer be accessible
♻ ☆ Structured Pruning Adapters
Adapters are a parameter-efficient alternative to fine-tuning, which augment a frozen base network to learn new tasks. Yet, the inference of the adapted model is often slower than the corresponding fine-tuned model. To improve on this, we propose Structured Pruning Adapters (SPAs), a family of compressing, task-switching network adapters, that accelerate and specialize networks using tiny parameter sets and structured pruning. Specifically, we propose a channel-based SPA and evaluate it with a suite of pruning methods on multiple computer vision benchmarks. Compared to regular structured pruning with fine-tuning, our channel-SPAs improve accuracy by 6.9% on average while using half the parameters at 90% pruned weights. Alternatively, they can learn adaptations with 17x fewer parameters at 70% pruning with 1.6% lower accuracy. Similarly, our block-SPA requires far fewer parameters than pruning with fine-tuning. Our experimental code and Python library of adapters are available at github.com/lukashedegaard/structured-pruning-adapters.
comment: 11 pages, 6 figures, 2 tables
♻ ☆ Towards Label-Efficient Incremental Learning: A Survey
The current dominant paradigm when building a machine learning model is to iterate over a dataset over and over until convergence. Such an approach is non-incremental, as it assumes access to all images of all categories at once. However, for many applications, non-incremental learning is unrealistic. To that end, researchers study incremental learning, where a learner is required to adapt to an incoming stream of data with a varying distribution while preventing forgetting of past knowledge. Significant progress has been made, however, the vast majority of works focus on the fully supervised setting, making these algorithms label-hungry thus limiting their real-life deployment. To that end, in this paper, we make the first attempt to survey recently growing interest in label-efficient incremental learning. We identify three subdivisions, namely semi-, few-shot- and self-supervised learning to reduce labeling efforts. Finally, we identify novel directions that can further enhance label-efficiency and improve incremental learning scalability. Project website: https://github.com/kilickaya/label-efficient-il.
♻ ☆ Are Labels Needed for Incremental Instance Learning?
In this paper, we learn to classify visual object instances, incrementally and via self-supervision (self-incremental). Our learner observes a single instance at a time, which is then discarded from the dataset. Incremental instance learning is challenging, since longer learning sessions exacerbate forgetfulness, and labeling instances is cumbersome. We overcome these challenges via three contributions: i. We propose VINIL, a self-incremental learner that can learn object instances sequentially, ii. We equip VINIL with self-supervision to by-pass the need for instance labelling, iii. We compare VINIL to label-supervised variants on two large-scale benchmarks and show that VINIL significantly improves accuracy while reducing forgetfulness.
♻ ☆ BRAIxDet: Learning to Detect Malignant Breast Lesion with Incomplete Annotations
Methods to detect malignant lesions from screening mammograms are usually trained with fully annotated datasets, where images are labelled with the localisation and classification of cancerous lesions. However, real-world screening mammogram datasets commonly have a subset that is fully annotated and another subset that is weakly annotated with just the global classification (i.e., without lesion localisation). Given the large size of such datasets, researchers usually face a dilemma with the weakly annotated subset: to not use it or to fully annotate it. The first option will reduce detection accuracy because it does not use the whole dataset, and the second option is too expensive given that the annotation needs to be done by expert radiologists. In this paper, we propose a middle-ground solution for the dilemma, which is to formulate the training as a weakly- and semi-supervised learning problem that we refer to as malignant breast lesion detection with incomplete annotations. To address this problem, our new method comprises two stages, namely: 1) pre-training a multi-view mammogram classifier with weak supervision from the whole dataset, and 2) extending the trained classifier to become a multi-view detector that is trained with semi-supervised student-teacher learning, where the training set contains fully and weakly-annotated mammograms. We provide extensive detection results on two real-world screening mammogram datasets containing incomplete annotations, and show that our proposed approach achieves state-of-the-art results in the detection of malignant breast lesions with incomplete annotations.
comment: Under Review
♻ ☆ Autoencoders as Cross-Modal Teachers: Can Pretrained 2D Image Transformers Help 3D Representation Learning? ICLR 2023
The success of deep learning heavily relies on large-scale data with comprehensive labels, which is more expensive and time-consuming to fetch in 3D compared to 2D images or natural languages. This promotes the potential of utilizing models pretrained with data more than 3D as teachers for cross-modal knowledge transferring. In this paper, we revisit masked modeling in a unified fashion of knowledge distillation, and we show that foundational Transformers pretrained with 2D images or natural languages can help self-supervised 3D representation learning through training Autoencoders as Cross-Modal Teachers (ACT). The pretrained Transformers are transferred as cross-modal 3D teachers using discrete variational autoencoding self-supervision, during which the Transformers are frozen with prompt tuning for better knowledge inheritance. The latent features encoded by the 3D teachers are used as the target of masked point modeling, wherein the dark knowledge is distilled to the 3D Transformer students as foundational geometry understanding. Our ACT pretrained 3D learner achieves state-of-the-art generalization capacity across various downstream benchmarks, e.g., 88.21% overall accuracy on ScanObjectNN. Codes have been released at https://github.com/RunpeiDong/ACT.
comment: Accepted at ICLR 2023
♻ ☆ SOTVerse: A User-defined Task Space of Single Object Tracking
Single object tracking (SOT) research falls into a cycle -- trackers perform well on most benchmarks but quickly fail in challenging scenarios, causing researchers to doubt the insufficient data content and take more effort to construct larger datasets with more challenging situations. However, inefficient data utilization and limited evaluation methods more seriously hinder SOT research. The former causes existing datasets can not be exploited comprehensively, while the latter neglects challenging factors in the evaluation process. In this article, we systematize the representative benchmarks and form a Single Object Tracking metaverse (SOTVerse) -- a user-defined SOT task space to break through the bottleneck. We first propose a 3E Paradigm to describe tasks by three components (i.e., environment, evaluation, and executor). Then, we summarize task characteristics, clarify the organization standards, and construct SOTVerse with 12.56 million frames. Specifically, SOTVerse automatically labels challenging factors per frame, allowing users to generate user-defined spaces efficiently via construction rules. Besides, SOTVerse provides two mechanisms with new indicators and successfully evaluates trackers under various subtasks. Consequently, SOTVerse first provides a strategy to improve resource utilization in the computer vision area, making research more standardized and scientific. The SOTVerse, toolkit, evaluation server, and results are available at http://metaverse.aitestunion.com.
comment: This paper is submitted to an international journal
♻ ☆ R2FD2: Fast and Robust Matching of Multimodal Remote Sensing Image via Repeatable Feature Detector and Rotation-invariant Feature Descriptor
Automatically identifying feature correspondences between multimodal images is facing enormous challenges because of the significant differences both in radiation and geometry. To address these problems, we propose a novel feature matching method (named R2FD2) that is robust to radiation and rotation differences. Our R2FD2 is conducted in two critical contributions, consisting of a repeatable feature detector and a rotation-invariant feature descriptor. In the first stage, a repeatable feature detector called the Multi-channel Auto-correlation of the Log-Gabor (MALG) is presented for feature detection, which combines the multi-channel auto-correlation strategy with the Log-Gabor wavelets to detect interest points (IPs) with high repeatability and uniform distribution. In the second stage, a rotation-invariant feature descriptor is constructed, named the Rotation-invariant Maximum index map of the Log-Gabor (RMLG), which consists of two components: fast assignment of dominant orientation and construction of feature representation. In the process of fast assignment of dominant orientation, a Rotation-invariant Maximum Index Map (RMIM) is built to address rotation deformations. Then, the proposed RMLG incorporates the rotation-invariant RMIM with the spatial configuration of DAISY to depict a more discriminative feature representation, which improves RMLG's resistance to radiation and rotation variances.Experimental results show that the proposed R2FD2 outperforms five state-of-the-art feature matching methods, and has superior advantages in adaptability and universality. Moreover, our R2FD2 achieves the accuracy of matching within two pixels and has a great advantage in matching efficiency over other state-of-the-art methods.
comment: 33 pages, 15 figures
♻ ☆ The KFIoU Loss for Rotated Object Detection ICLR 2023
Differing from the well-developed horizontal object detection area whereby the computing-friendly IoU based loss is readily adopted and well fits with the detection metrics. In contrast, rotation detectors often involve a more complicated loss based on SkewIoU which is unfriendly to gradient-based training. In this paper, we propose an effective approximate SkewIoU loss based on Gaussian modeing and Kalman filter, which mainly consists of two items. The first term is a scale-insensitive center point loss, which is used to quickly get the center points between bounding boxes closer to assist the second term. In the distance-independent second term, Kalman filter is adopted to inherently mimic the mechanism of SkewIoU by its definition, and show its alignment with the SkewIoU loss at trend-level within a certain distance (i.e. within 9 pixels). This is in contrast to recent Gaussian modeling based rotation detectors e.g. GWD loss and KLD loss that involve a human-specified distribution distance metric which require additional hyperparameter tuning that vary across datasets and detectors. The resulting new loss called KFIoU loss is easier to implement and works better compared with exact SkewIoU loss, thanks to its full differentiability and ability to handle the non-overlapping cases. We further extend our technique to the 3-D case which also suffers from the same issues as 2-D detection. Extensive results on various public datasets (2-D/3-D, aerial/text/face images) with different base detectors show the effectiveness of our approach.
comment: 17 pages, 6 figures, 7 tables, accepted by ICLR 2023, TensorFlow code: https://github.com/yangxue0827/RotationDetection, PyTorch code: https://github.com/open-mmlab/mmrotate, Jittor code: https://github.com/Jittor/JDet
♻ ☆ H2RBox: Horizontal Box Annotation is All You Need for Oriented Object Detection ICLR 2023
Oriented object detection emerges in many applications from aerial images to autonomous driving, while many existing detection benchmarks are annotated with horizontal bounding box only which is also less costive than fine-grained rotated box, leading to a gap between the readily available training corpus and the rising demand for oriented object detection. This paper proposes a simple yet effective oriented object detection approach called H2RBox merely using horizontal box annotation for weakly-supervised training, which closes the above gap and shows competitive performance even against those trained with rotated boxes. The cores of our method are weakly- and self-supervised learning, which predicts the angle of the object by learning the consistency of two different views. To our best knowledge, H2RBox is the first horizontal box annotation-based oriented object detector. Compared to an alternative i.e. horizontal box-supervised instance segmentation with our post adaption to oriented object detection, our approach is not susceptible to the prediction quality of mask and can perform more robustly in complex scenes containing a large number of dense objects and outliers. Experimental results show that H2RBox has significant performance and speed advantages over horizontal box-supervised instance segmentation methods, as well as lower memory requirements. While compared to rotated box-supervised oriented object detectors, our method shows very close performance and speed. The source code is available at https://github.com/yangxue0827/h2rbox-mmrotate and https://github.com/yangxue0827/h2rbox-jittor.
comment: 15 pages, 6 figures, 7 tables, accepted by ICLR 2023, the source code is available at https://github.com/yangxue0827/h2rbox-mmrotate and https://github.com/yangxue0827/h2rbox-jittor
♻ ☆ The Value of Out-of-Distribution Data ECCV 2022
We expect the generalization error to improve with more samples from a similar task, and to deteriorate with more samples from an out-of-distribution (OOD) task. In this work, we show a counter-intuitive phenomenon: the generalization error of a task can be a non-monotonic function of the number of OOD samples. As the number of OOD samples increases, the generalization error on the target task improves before deteriorating beyond a threshold. In other words, there is value in training on small amounts of OOD data. We use Fisher's Linear Discriminant on synthetic datasets and deep networks on computer vision benchmarks such as MNIST, CIFAR-10, CINIC-10, PACS and DomainNet to demonstrate and analyze this phenomenon. In the idealistic setting where we know which samples are OOD, we show that these non-monotonic trends can be exploited using an appropriately weighted objective of the target and OOD empirical risk. While its practical utility is limited, this does suggest that if we can detect OOD samples, then there may be ways to benefit from them. When we do not know which samples are OOD, we show how a number of go-to strategies such as data-augmentation, hyper-parameter optimization, and pre-training are not enough to ensure that the target generalization error does not deteriorate with the number of OOD samples in the dataset.
comment: Previous versions of this work have been presented at the Out-of-Distribution Generalization in Computer Vision (OOD-CV) Workshop (ECCV 2022) and the Workshop on Distribution Shifts (NeurIPS 2022)
♻ ☆ Adversarially Robust Neural Architectures
Deep Neural Networks (DNNs) are vulnerable to adversarial attacks. Existing methods are devoted to developing various robust training strategies or regularizations to update the weights of the neural network. But beyond the weights, the overall structure and information flow in the network are explicitly determined by the neural architecture, which remains unexplored. This paper thus aims to improve the adversarial robustness of the network from the architecture perspective. We explore the relationship among adversarial robustness, Lipschitz constant, and architecture parameters and show that an appropriate constraint on architecture parameters could reduce the Lipschitz constant to further improve the robustness. The importance of architecture parameters could vary from operation to operation or connection to connection. We approximate the Lipschitz constant of the entire network through a univariate log-normal distribution, whose mean and variance are related to architecture parameters. The confidence can be fulfilled through formulating a constraint on the distribution parameters based on the cumulative function. Compared with adversarially trained neural architectures searched by various NAS algorithms as well as efficient human-designed models, our algorithm empirically achieves the best performance among all the models under various attacks on different datasets.
comment: 13 pages, 5 figures, 8 tables
♻ ☆ 3D Reconstruction of Non-cooperative Resident Space Objects using Instant NGP-accelerated NeRF and D-NeRF
The proliferation of non-cooperative resident space objects (RSOs) in orbit has spurred the demand for active space debris removal, on-orbit servicing (OOS), classification, and functionality identification of these RSOs. Recent advances in computer vision have enabled high-definition 3D modeling of objects based on a set of 2D images captured from different viewing angles. This work adapts Instant NeRF and D-NeRF, variations of the neural radiance field (NeRF) algorithm to the problem of mapping RSOs in orbit for the purposes of functionality identification and assisting with OOS. The algorithms are evaluated for 3D reconstruction quality and hardware requirements using datasets of images of a spacecraft mock-up taken under two different lighting and motion conditions at the Orbital Robotic Interaction, On-Orbit Servicing and Navigation (ORION) Laboratory at Florida Institute of Technology. Instant NeRF is shown to learn high-fidelity 3D models with a computational cost that could feasibly be trained on on-board computers.
comment: Presented at AAS/AIAA Spaceflight Mechanics Conference 2023, 14 pages, 10 figures, 2 tables
♻ ☆ HALOC: Hardware-Aware Automatic Low-Rank Compression for Compact Neural Networks AAAI-23
Low-rank compression is an important model compression strategy for obtaining compact neural network models. In general, because the rank values directly determine the model complexity and model accuracy, proper selection of layer-wise rank is very critical and desired. To date, though many low-rank compression approaches, either selecting the ranks in a manual or automatic way, have been proposed, they suffer from costly manual trials or unsatisfied compression performance. In addition, all of the existing works are not designed in a hardware-aware way, limiting the practical performance of the compressed models on real-world hardware platforms. To address these challenges, in this paper we propose HALOC, a hardware-aware automatic low-rank compression framework. By interpreting automatic rank selection from an architecture search perspective, we develop an end-to-end solution to determine the suitable layer-wise ranks in a differentiable and hardware-aware way. We further propose design principles and mitigation strategy to efficiently explore the rank space and reduce the potential interference problem. Experimental results on different datasets and hardware platforms demonstrate the effectiveness of our proposed approach. On CIFAR-10 dataset, HALOC enables 0.07% and 0.38% accuracy increase over the uncompressed ResNet-20 and VGG-16 models with 72.20% and 86.44% fewer FLOPs, respectively. On ImageNet dataset, HALOC achieves 0.9% higher top-1 accuracy than the original ResNet-18 model with 66.16% fewer FLOPs. HALOC also shows 0.66% higher top-1 accuracy increase than the state-of-the-art automatic low-rank compression solution with fewer computational and memory costs. In addition, HALOC demonstrates the practical speedups on different hardware platforms, verified by the measurement results on desktop GPU, embedded GPU and ASIC accelerator.
comment: AAAI-23
♻ ☆ Beyond Invariance: Test-Time Label-Shift Adaptation for Distributions with "Spurious" Correlations
Spurious correlations, or correlations that change across domains where a model can be deployed, present significant challenges to real-world applications of machine learning models. However, such correlations are not always "spurious"; often, they provide valuable prior information for a prediction. Here, we present a test-time adaptation method that exploits the spurious correlation phenomenon, in contrast to recent approaches that attempt to eliminate spurious correlations through invariance. We consider situations where the prior distribution $p(y, z)$, which models the dependence between the class label $y$ and the "nuisance" factors $z$, may change across domains, but the generative model for features $p(\mathbf{x}|y, z)$ is constant. We note that this corresponds to an expanded version of the label shift assumption, where the labels now also include the nuisance factors $z$. Based on this observation, we train a classifier to predict $p(y, z|\mathbf{x})$ on the source distribution, and propose a test-time label shift correction that adapts to changes in the marginal distribution $p(y, z)$ using unlabeled samples from the target domain. We evaluate our method, which we call "Test-Time Label-Shift Adaptation" (TTLSA), on two different image datasets -- the CheXpert chest X-ray dataset and the Colored MNIST dataset -- and show a significant improvement over baseline methods. Code reproducing experiments is available at https://github.com/nalzok/test-time-label-shift .
comment: 17 pages, 7 figures
♻ ☆ Few-Shot Learning Enables Population-Scale Analysis of Leaf Traits in Populus trichocarpa
Plant phenotyping is typically a time-consuming and expensive endeavor, requiring large groups of researchers to meticulously measure biologically relevant plant traits, and is the main bottleneck in understanding plant adaptation and the genetic architecture underlying complex traits at population scale. In this work, we address these challenges by leveraging few-shot learning with convolutional neural networks (CNNs) to segment the leaf body and visible venation of 2,906 P. trichocarpa leaf images obtained in the field. In contrast to previous methods, our approach (i) does not require experimental or image pre-processing, (ii) uses the raw RGB images at full resolution, and (iii) requires very few samples for training (e.g., just eight images for vein segmentation). Traits relating to leaf morphology and vein topology are extracted from the resulting segmentations using traditional open-source image-processing tools, validated using real-world physical measurements, and used to conduct a genome-wide association study to identify genes controlling the traits. In this way, the current work is designed to provide the plant phenotyping community with (i) methods for fast and accurate image-based feature extraction that require minimal training data, and (ii) a new population-scale data set, including 68 different leaf phenotypes, for domain scientists and machine learning researchers. All of the few-shot learning code, data, and results are made publicly available.
♻ ☆ FiT: Parameter Efficient Few-shot Transfer Learning for Personalized and Federated Image Classification
Modern deep learning systems are increasingly deployed in situations such as personalization and federated learning where it is necessary to support i) learning on small amounts of data, and ii) communication efficient distributed training protocols. In this work, we develop FiLM Transfer (FiT) which fulfills these requirements in the image classification setting by combining ideas from transfer learning (fixed pretrained backbones and fine-tuned FiLM adapter layers) and meta-learning (automatically configured Naive Bayes classifiers and episodic training) to yield parameter efficient models with superior classification accuracy at low-shot. The resulting parameter efficiency is key for enabling few-shot learning, inexpensive model updates for personalization, and communication efficient federated learning. We experiment with FiT on a wide range of downstream datasets and show that it achieves better classification accuracy than the leading Big Transfer (BiT) algorithm at low-shot and achieves state-of-the art accuracy on the challenging VTAB-1k benchmark, with fewer than 1% of the updateable parameters. Finally, we demonstrate the parameter efficiency and superior accuracy of FiT in distributed low-shot applications including model personalization and federated learning where model update size is an important performance metric.
Information Retrieval 7
☆ Practical Bandits: An Industry Perspective WWW
The bandit paradigm provides a unified modeling framework for problems that require decision-making under uncertainty. Because many business metrics can be viewed as rewards (a.k.a. utilities) that result from actions, bandit algorithms have seen a large and growing interest from industrial applications, such as search, recommendation and advertising. Indeed, with the bandit lens comes the promise of direct optimisation for the metrics we care about. Nevertheless, the road to successfully applying bandits in production is not an easy one. Even when the action space and rewards are well-defined, practitioners still need to make decisions regarding multi-arm or contextual approaches, on- or off-policy setups, delayed or immediate feedback, myopic or long-term optimisation, etc. To make matters worse, industrial platforms typically give rise to large action spaces in which existing approaches tend to break down. The research literature on these topics is broad and vast, but this can overwhelm practitioners, whose primary aim is to solve practical problems, and therefore need to decide on a specific instantiation or approach for each project. This tutorial will take a step towards filling that gap between the theory and practice of bandits. Our goal is to present a unified overview of the field and its existing terminology, concepts and algorithms -- with a focus on problems relevant to industry. We hope our industrial perspective will help future practitioners who wish to leverage the bandit paradigm for their application.
comment: Tutorial held at The Web Conference 2023 (formerly known as WWW) in Austin, Texas (USA), on April 30 - May 4, 2023
☆ PEPNet: Parameter and Embedding Personalized Network for Infusing with Personalized Prior Information
With the increase of pages and buttons in real-world applications, industrial-scale recommender systems face multi-domain and multi-task challenges. On the one hand, users and items in multiple domains suffer inconsistent distributions. On the other hand, multiple tasks have distinctive sparsity and interdependence. Personalization modeling is the core of recommender systems. Accurate personalization estimation helps to capture the degree of user preference for items in different situations, especially in the case of multiple domains and multiple tasks. In multi-task and multi-domain recommendation, how to introduce personalized priors into the model in the right place and in the right way is crucial. In this paper, we propose a plug-and-play Parameter and Embedding Personalized Network (PEPNet) for multi-task recommendation in the multi-domain setting. PEPNet takes features with strong bias as input and dynamically acts on the bottom-layer embeddings or the top-layer DNN hidden units in the model through the gate mechanism. By mapping significant priors to scaling weights ranging from 0 to 2, PEPNet introduces both parameter personalization and embedding personalization. Embedding Personalized Network (EPNet) selects and aligns embeddings with different semantics under multiple domains. Parameter Personalized Network (PPNet) influences DNN parameters to balance interdependent targets in multiple tasks. To further adapt to the characteristics of the model, we have made corresponding engineering optimizations on the Embedding and DNN parameter update strategies. We have deployed the model in Kuaishou and Kuaishou Express apps, serving over 300 million daily users. Both online and offline experiments have demonstrated substantial improvements in multiple metrics. In particular, we have seen a more than 1\% online increase in three major domains.
☆ Goniometers are a Powerful Acoustic Feature for Music Information Retrieval Tasks
Goniometers, also known as Phase Scopes or Vector Scopes, are audio metering tools that help music producers and mixing engineers monitor spatial aspects of a music mix, such as the stereo panorama, the width of single sources, the amount and diffuseness of reverberation as well as phase cancellations that may occur on the sweet-spot and in a mono-mixdown. In addition, they implicitly inform about the dynamics of the sound. Self-organizing maps trained with a goniometer, are consulted to explore the usefulness of this acoustic feature for music information retrieval tasks. One can see that goniometers are able to classify different genres and cluster a single album. The advantage of goniometers is the causality: Music producers and mixing engineers consciously consult goniometers to reach their desired sound, which is not the case for other acoustic features, from Zero-Crossing Rate to Mel-Frequency Cepstral Coefficients.
☆ MMRec: Simplifying Multimodal Recommendation
This paper presents an open-source toolbox, MMRec for multimodal recommendation. MMRec simplifies and canonicalizes the process of implementing and comparing multimodal recommendation models. The objective of MMRec is to provide a unified and configurable arena that can minimize the effort in implementing and testing multimodal recommendation models. It enables multimodal models, ranging from traditional matrix factorization to modern graph-based algorithms, capable of fusing information from multiple modalities simultaneously. Our documentation, examples, and source code are available at \url{https://github.com/enoche/MMRec}.
comment: 3 pages
♻ ☆ Large-scale Stochastic Optimization of NDCG Surrogates for Deep Learning with Provable Convergence ICML2022
NDCG, namely Normalized Discounted Cumulative Gain, is a widely used ranking metric in information retrieval and machine learning. However, efficient and provable stochastic methods for maximizing NDCG are still lacking, especially for deep models. In this paper, we propose a principled approach to optimize NDCG and its top-$K$ variant. First, we formulate a novel compositional optimization problem for optimizing the NDCG surrogate, and a novel bilevel compositional optimization problem for optimizing the top-$K$ NDCG surrogate. Then, we develop efficient stochastic algorithms with provable convergence guarantees for the non-convex objectives. Different from existing NDCG optimization methods, the per-iteration complexity of our algorithms scales with the mini-batch size instead of the number of total items. To improve the effectiveness for deep learning, we further propose practical strategies by using initial warm-up and stop gradient operator. Experimental results on multiple datasets demonstrate that our methods outperform prior ranking approaches in terms of NDCG. To the best of our knowledge, this is the first time that stochastic algorithms are proposed to optimize NDCG with a provable convergence guarantee. Our proposed methods are implemented in the LibAUC library at https://libauc.org/.
comment: 32 pages, 12 figures; Accepted by ICML2022
♻ ☆ Is it worth it? Comparing six deep and classical methods for unsupervised anomaly detection in time series
Detecting anomalies in time series data is important in a variety of fields, including system monitoring, healthcare, and cybersecurity. While the abundance of available methods makes it difficult to choose the most appropriate method for a given application, each method has its strengths in detecting certain types of anomalies. In this study, we compare six unsupervised anomaly detection methods of varying complexity to determine whether more complex methods generally perform better and if certain methods are better suited to certain types of anomalies. We evaluated the methods using the UCR anomaly archive, a recent benchmark dataset for anomaly detection. We analyzed the results on a dataset and anomaly type level after adjusting the necessary hyperparameters for each method. Additionally, we assessed the ability of each method to incorporate prior knowledge about anomalies and examined the differences between point-wise and sequence-wise features. Our experiments show that classical machine learning methods generally outperform deep learning methods across a range of anomaly types.
comment: 17 Pages, The repository to reproduce the results is available at https://gitlab.com/dlr-dw/is-it-worth-it-benchmark
♻ ☆ Knowledge Extraction in Low-Resource Scenarios: Survey and Perspective
Knowledge Extraction (KE), aiming to extract structural information from unstructured texts, often suffers from data scarcity and emerging unseen types, i.e., low-resource scenarios. Many neural approaches to low-resource KE have been widely investigated and achieved impressive performance. In this paper, we present a literature review towards KE in low-resource scenarios, and systematically categorize existing works into three paradigms: (1) exploiting higher-resource data, (2) exploiting stronger models, and (3) exploiting data and models together. In addition, we highlight promising applications and outline some potential directions for future research. We hope that our survey can help both the academic and industrial communities to better understand this field, inspire more ideas, and boost broader applications.
comment: Work in Progress
Machine Learning 168
☆ Lower Bounds for Learning in Revealing POMDPs
This paper studies the fundamental limits of reinforcement learning (RL) in the challenging \emph{partially observable} setting. While it is well-established that learning in Partially Observable Markov Decision Processes (POMDPs) requires exponentially many samples in the worst case, a surge of recent work shows that polynomial sample complexities are achievable under the \emph{revealing condition} -- A natural condition that requires the observables to reveal some information about the unobserved latent states. However, the fundamental limits for learning in revealing POMDPs are much less understood, with existing lower bounds being rather preliminary and having substantial gaps from the current best upper bounds. We establish strong PAC and regret lower bounds for learning in revealing POMDPs. Our lower bounds scale polynomially in all relevant problem parameters in a multiplicative fashion, and achieve significantly smaller gaps against the current best upper bounds, providing a solid starting point for future studies. In particular, for \emph{multi-step} revealing POMDPs, we show that (1) the latent state-space dependence is at least $\Omega(S^{1.5})$ in the PAC sample complexity, which is notably harder than the $\widetilde{\Theta}(S)$ scaling for fully-observable MDPs; (2) Any polynomial sublinear regret is at least $\Omega(T^{2/3})$, suggesting its fundamental difference from the \emph{single-step} case where $\widetilde{O}(\sqrt{T})$ regret is achievable. Technically, our hard instance construction adapts techniques in \emph{distribution testing}, which is new to the RL literature and may be of independent interest.
☆ Bayesian Metric Learning for Uncertainty Quantification in Image Retrieval
We propose the first Bayesian encoder for metric learning. Rather than relying on neural amortization as done in prior works, we learn a distribution over the network weights with the Laplace Approximation. We actualize this by first proving that the contrastive loss is a valid log-posterior. We then propose three methods that ensure a positive definite Hessian. Lastly, we present a novel decomposition of the Generalized Gauss-Newton approximation. Empirically, we show that our Laplacian Metric Learner (LAM) estimates well-calibrated uncertainties, reliably detects out-of-distribution examples, and yields state-of-the-art predictive performance.
comment: Code: https://github.com/FrederikWarburg/bayesian-metric-learning
☆ $IC^3$: Image Captioning by Committee Consensus
If you ask a human to describe an image, they might do so in a thousand different ways. Traditionally, image captioning models are trained to approximate the reference distribution of image captions, however, doing so encourages captions that are viewpoint-impoverished. Such captions often focus on only a subset of the possible details, while ignoring potentially useful information in the scene. In this work, we introduce a simple, yet novel, method: "Image Captioning by Committee Consensus" ($IC^3$), designed to generate a single caption that captures high-level details from several viewpoints. Notably, humans rate captions produced by $IC^3$ at least as helpful as baseline SOTA models more than two thirds of the time, and $IC^3$ captions can improve the performance of SOTA automated recall systems by up to 84%, indicating significant material improvements over existing SOTA approaches for visual description. Our code is publicly available at https://github.com/DavidMChan/caption-by-committee
☆ Dual PatchNorm
We propose Dual PatchNorm: two Layer Normalization layers (LayerNorms), before and after the patch embedding layer in Vision Transformers. We demonstrate that Dual PatchNorm outperforms the result of exhaustive search for alternative LayerNorm placement strategies in the Transformer block itself. In our experiments, incorporating this trivial modification, often leads to improved accuracy over well-tuned Vision Transformers and never hurts.
☆ Federated Analytics: A survey
Federated analytics (FA) is a privacy-preserving framework for computing data analytics over multiple remote parties (e.g., mobile devices) or silo-ed institutional entities (e.g., hospitals, banks) without sharing the data among parties. Motivated by the practical use cases of federated analytics, we follow a systematic discussion on federated analytics in this article. In particular, we discuss the unique characteristics of federated analytics and how it differs from federated learning. We also explore a wide range of FA queries and discuss various existing solutions and potential use case applications for different FA queries.
comment: To appear in APSIPA Transactions on Signal and Information Processing, Volume 12, Issue 1
☆ Randomized Greedy Learning for Non-monotone Stochastic Submodular Maximization Under Full-bandit Feedback
We investigate the problem of unconstrained combinatorial multi-armed bandits with full-bandit feedback and stochastic rewards for submodular maximization. Previous works investigate the same problem assuming a submodular and monotone reward function. In this work, we study a more general problem, i.e., when the reward function is not necessarily monotone, and the submodularity is assumed only in expectation. We propose Randomized Greedy Learning (RGL) algorithm and theoretically prove that it achieves a $\frac{1}{2}$-regret upper bound of $\tilde{\mathcal{O}}(n T^{\frac{2}{3}})$ for horizon $T$ and number of arms $n$. We also show in experiments that RGL empirically outperforms other full-bandit variants in submodular and non-submodular settings.
☆ Are Diffusion Models Vulnerable to Membership Inference Attacks?
Diffusion-based generative models have shown great potential for image synthesis, but there is a lack of research on the security and privacy risks they may pose. In this paper, we investigate the vulnerability of diffusion models to Membership Inference Attacks (MIAs), a common privacy concern. Our results indicate that existing MIAs designed for GANs or VAE are largely ineffective on diffusion models, either due to inapplicable scenarios (e.g., requiring the discriminator of GANs) or inappropriate assumptions (e.g., closer distances between synthetic images and member images). To address this gap, we propose Step-wise Error Comparing Membership Inference (SecMI), a black-box MIA that infers memberships by assessing the matching of forward process posterior estimation at each timestep. SecMI follows the common overfitting assumption in MIA where member samples normally have smaller estimation errors, compared with hold-out samples. We consider both the standard diffusion models, e.g., DDPM, and the text-to-image diffusion models, e.g., Stable Diffusion. Experimental results demonstrate that our methods precisely infer the membership with high confidence on both of the two scenarios across six different datasets
☆ Double Permutation Equivariance for Knowledge Graph Completion
This work provides a formalization of Knowledge Graphs (KGs) as a new class of graphs that we denote doubly exchangeable attributed graphs, where node and pairwise (joint 2-node) representations must be equivariant to permutations of both node ids and edge (& node) attributes (relations & node features). Double-permutation equivariant KG representations open a new research direction in KGs. We show that this equivariance imposes a structural representation of relations that allows neural networks to perform complex logical reasoning tasks in KGs. Finally, we introduce a general blueprint for such equivariant representations and test a simple GNN-based double-permutation equivariant neural architecture that achieve 100% Hits@10 test accuracy in both the WN18RRv1 and NELL995v1 inductive KG completion tasks, and can accurately perform logical reasoning tasks that no existing methods can perform, to the best of our knowledge.
☆ Normalizing Flow Ensembles for Rich Aleatoric and Epistemic Uncertainty Modeling
In this work, we demonstrate how to reliably estimate epistemic uncertainty while maintaining the flexibility needed to capture complicated aleatoric distributions. To this end, we propose an ensemble of Normalizing Flows (NF), which are state-of-the-art in modeling aleatoric uncertainty. The ensembles are created via sets of fixed dropout masks, making them less expensive than creating separate NF models. We demonstrate how to leverage the unique structure of NFs, base distributions, to estimate aleatoric uncertainty without relying on samples, provide a comprehensive set of baselines, and derive unbiased estimates for differential entropy. The methods were applied to a variety of experiments, commonly used to benchmark aleatoric and epistemic uncertainty estimation: 1D sinusoidal data, 2D windy grid-world ($\it{Wet Chicken}$), $\it{Pendulum}$, and $\it{Hopper}$. In these experiments, we setup an active learning framework and evaluate each model's capability at measuring aleatoric and epistemic uncertainty. The results show the advantages of using NF ensembles in capturing complicated aleatoric while maintaining accurate epistemic uncertainty estimates.
☆ Bayesian Optimization of Multiple Objectives with Different Latencies
Multi-objective Bayesian optimization aims to find the Pareto front of optimal trade-offs between a set of expensive objectives while collecting as few samples as possible. In some cases, it is possible to evaluate the objectives separately, and a different latency or evaluation cost can be associated with each objective. This presents an opportunity to learn the Pareto front faster by evaluating the cheaper objectives more frequently. We propose a scalarization based knowledge gradient acquisition function which accounts for the different evaluation costs of the objectives. We prove consistency of the algorithm and show empirically that it significantly outperforms a benchmark algorithm which always evaluates both objectives.
comment: 25 pages
☆ What Language Reveals about Perception: Distilling Psychophysical Knowledge from Large Language Models
Understanding the extent to which the perceptual world can be recovered from language is a fundamental problem in cognitive science. We reformulate this problem as that of distilling psychophysical information from text and show how this can be done by combining large language models (LLMs) with a classic psychophysical method based on similarity judgments. Specifically, we use the prompt auto-completion functionality of GPT3, a state-of-the-art LLM, to elicit similarity scores between stimuli and then apply multidimensional scaling to uncover their underlying psychological space. We test our approach on six perceptual domains and show that the elicited judgments strongly correlate with human data and successfully recover well-known psychophysical structures such as the color wheel and pitch spiral. We also explore meaningful divergences between LLM and human representations. Our work showcases how combining state-of-the-art machine models with well-known cognitive paradigms can shed new light on fundamental questions in perception and language research.
comment: 7 pages, 5 figures
☆ Bayesian Inference on Binary Spiking Networks Leveraging Nanoscale Device Stochasticity ISCA
Bayesian Neural Networks (BNNs) can overcome the problem of overconfidence that plagues traditional frequentist deep neural networks, and are hence considered to be a key enabler for reliable AI systems. However, conventional hardware realizations of BNNs are resource intensive, requiring the implementation of random number generators for synaptic sampling. Owing to their inherent stochasticity during programming and read operations, nanoscale memristive devices can be directly leveraged for sampling, without the need for additional hardware resources. In this paper, we introduce a novel Phase Change Memory (PCM)-based hardware implementation for BNNs with binary synapses. The proposed architecture consists of separate weight and noise planes, in which PCM cells are configured and operated to represent the nominal values of weights and to generate the required noise for sampling, respectively. Using experimentally observed PCM noise characteristics, for the exemplary Breast Cancer Dataset classification problem, we obtain hardware accuracy and expected calibration error matching that of an 8-bit fixed-point (FxP8) implementation, with projected savings of over 9$\times$ in terms of core area transistor count.
comment: Submitted and Accepted in ISCAS 2023
☆ MARLIN: Soft Actor-Critic based Reinforcement Learning for Congestion Control in Real Networks AAAI 2023
Fast and efficient transport protocols are the foundation of an increasingly distributed world. The burden of continuously delivering improved communication performance to support next-generation applications and services, combined with the increasing heterogeneity of systems and network technologies, has promoted the design of Congestion Control (CC) algorithms that perform well under specific environments. The challenge of designing a generic CC algorithm that can adapt to a broad range of scenarios is still an open research question. To tackle this challenge, we propose to apply a novel Reinforcement Learning (RL) approach. Our solution, MARLIN, uses the Soft Actor-Critic algorithm to maximize both entropy and return and models the learning process as an infinite-horizon task. We trained MARLIN on a real network with varying background traffic patterns to overcome the sim-to-real mismatch that researchers have encountered when applying RL to CC. We evaluated our solution on the task of file transfer and compared it to TCP Cubic. While further research is required, results have shown that MARLIN can achieve comparable results to TCP with little hyperparameter tuning, in a task significantly different from its training setting. Therefore, we believe that our work represents a promising first step toward building CC algorithms based on the maximum entropy RL framework.
comment: 10 pages, 5 figures, AAAI 2023 workshop "Reinforcement Learning Ready for Production", accepted at NOMS 2023 - IEEE/IFIP Network Operations and Management Symposium
☆ Convolutional Autoencoders, Clustering and POD for Low-dimensional Parametrization of Navier-Stokes Equations
Simulations of large-scale dynamical systems require expensive computations. Low-dimensional parametrization of high-dimensional states such as Proper Orthogonal Decomposition (POD) can be a solution to lessen the burdens by providing a certain compromise between accuracy and model complexity. However, for really low-dimensional parametrizations (for example for controller design) linear methods like the POD come to their natural limits so that nonlinear approaches will be the methods of choice. In this work we propose a convolutional autoencoder (CAE) consisting of a nonlinear encoder and an affine linear decoder and consider combinations with k-means clustering for improved encoding performance. The proposed set of methods is compared to the standard POD approach in two cylinder-wake scenarios modeled by the incompressible Navier-Stokes equations.
comment: 16 pages, 12 figures
☆ ReLOAD: Reinforcement Learning with Optimistic Ascent-Descent for Last-Iterate Convergence in Constrained MDPs
In recent years, Reinforcement Learning (RL) has been applied to real-world problems with increasing success. Such applications often require to put constraints on the agent's behavior. Existing algorithms for constrained RL (CRL) rely on gradient descent-ascent, but this approach comes with a caveat. While these algorithms are guaranteed to converge on average, they do not guarantee last-iterate convergence, i.e., the current policy of the agent may never converge to the optimal solution. In practice, it is often observed that the policy alternates between satisfying the constraints and maximizing the reward, rarely accomplishing both objectives simultaneously. Here, we address this problem by introducing Reinforcement Learning with Optimistic Ascent-Descent (ReLOAD), a principled CRL method with guaranteed last-iterate convergence. We demonstrate its empirical effectiveness on a wide variety of CRL problems including discrete MDPs and continuous control. In the process we establish a benchmark of challenging CRL problems.
☆ Geometric Deep Learning for Autonomous Driving: Unlocking the Power of Graph Neural Networks With CommonRoad-Geometric
Heterogeneous graphs offer powerful data representations for traffic, given their ability to model the complex interaction effects among a varying number of traffic participants and the underlying road infrastructure. With the recent advent of graph neural networks (GNNs) as the accompanying deep learning framework, the graph structure can be efficiently leveraged for various machine learning applications such as trajectory prediction. As a first of its kind, our proposed Python framework offers an easy-to-use and fully customizable data processing pipeline to extract standardized graph datasets from traffic scenarios. Providing a platform for GNN-based autonomous driving research, it improves comparability between approaches and allows researchers to focus on model implementation instead of dataset curation.
☆ adSformers: Personalization from Short-Term Sequences and Diversity of Representations in Etsy Ads
In this article, we present our approach to personalizing Etsy Ads through encoding and learning from short-term (one-hour) sequences of user actions and diverse representations. To this end we introduce a three-component adSformer diversifiable personalization module (ADPM) and illustrate how we use this module to derive a short-term dynamic user representation and personalize the Click-Through Rate (CTR) and Post-Click Conversion Rate (PCCVR) models used in sponsored search (ad) ranking. The first component of the ADPM is a custom transformer encoder that learns the inherent structure from the sequence of actions. ADPM's second component enriches the signal through visual, multimodal and textual pretrained representations. Lastly, the third ADPM component includes a "learned" on the fly average pooled representation. The ADPM-personalized CTR and PCCVR models, henceforth referred to as adSformer CTR and adSformer PCCVR, outperform the CTR and PCCVR production baselines by $+6.65\%$ and $+12.70\%$, respectively, in offline Precision-Recall Area Under the Curve (PR AUC). At the time of this writing, following the online gains in A/B tests, such as $+5.34\%$ in return on ad spend, a seller success metric, we are ramping up the adSformers to $100\%$ traffic in Etsy Ads.
☆ Avoiding Model Estimation in Robust Markov Decision Processes with a Generative Model
Robust Markov Decision Processes (MDPs) are getting more attention for learning a robust policy which is less sensitive to environment changes. There are an increasing number of works analyzing sample-efficiency of robust MDPs. However, most works study robust MDPs in a model-based regime, where the transition probability needs to be estimated and requires $\mathcal{O}(|\mathcal{S}|^2|\mathcal{A}|)$ storage in memory. A common way to solve robust MDPs is to formulate them as a distributionally robust optimization (DRO) problem. However, solving a DRO problem is non-trivial, so prior works typically assume a strong oracle to obtain the optimal solution of the DRO problem easily. To remove the need for an oracle, we first transform the original robust MDPs into an alternative form, as the alternative form allows us to use stochastic gradient methods to solve the robust MDPs. Moreover, we prove the alternative form still preserves the role of robustness. With this new formulation, we devise a sample-efficient algorithm to solve the robust MDPs in a model-free regime, from which we benefit lower memory space $\mathcal{O}(|\mathcal{S}||\mathcal{A}|)$ without using the oracle. Finally, we validate our theoretical findings via numerical experiments and show the efficiency to solve the alternative form of robust MDPs.
☆ Is Model Ensemble Necessary? Model-based RL via a Single Model with Lipschitz Regularized Value Function ICLR 2023
Probabilistic dynamics model ensemble is widely used in existing model-based reinforcement learning methods as it outperforms a single dynamics model in both asymptotic performance and sample efficiency. In this paper, we provide both practical and theoretical insights on the empirical success of the probabilistic dynamics model ensemble through the lens of Lipschitz continuity. We find that, for a value function, the stronger the Lipschitz condition is, the smaller the gap between the true dynamics- and learned dynamics-induced Bellman operators is, thus enabling the converged value function to be closer to the optimal value function. Hence, we hypothesize that the key functionality of the probabilistic dynamics model ensemble is to regularize the Lipschitz condition of the value function using generated samples. To test this hypothesis, we devise two practical robust training mechanisms through computing the adversarial noise and regularizing the value network's spectral norm to directly regularize the Lipschitz condition of the value functions. Empirical results show that combined with our mechanisms, model-based RL algorithms with a single dynamics model outperform those with an ensemble of probabilistic dynamics models. These findings not only support the theoretical insight, but also provide a practical solution for developing computationally efficient model-based RL algorithms.
comment: ICLR 2023
☆ Human not in the loop: objective sample difficulty measures for Curriculum Learning
Curriculum learning is a learning method that trains models in a meaningful order from easier to harder samples. A key here is to devise automatic and objective difficulty measures of samples. In the medical domain, previous work applied domain knowledge from human experts to qualitatively assess classification difficulty of medical images to guide curriculum learning, which requires extra annotation efforts, relies on subjective human experience, and may introduce bias. In this work, we propose a new automated curriculum learning technique using the variance of gradients (VoG) to compute an objective difficulty measure of samples and evaluated its effects on elbow fracture classification from X-ray images. Specifically, we used VoG as a metric to rank each sample in terms of the classification difficulty, where high VoG scores indicate more difficult cases for classification, to guide the curriculum training process We compared the proposed technique to a baseline (without curriculum learning), a previous method that used human annotations on classification difficulty, and anti-curriculum learning. Our experiment results showed comparable and higher performance for the binary and multi-class bone fracture classification tasks.
comment: ISBI 2023
☆ Neuro Symbolic Continual Learning: Knowledge, Reasoning Shortcuts and Concept Rehearsal
We introduce Neuro-Symbolic Continual Learning, where a model has to solve a sequence of neuro-symbolic tasks, that is, it has to map sub-symbolic inputs to high-level concepts and compute predictions by reasoning consistently with prior knowledge. Our key observation is that neuro-symbolic tasks, although different, often share concepts whose semantics remains stable over time. Traditional approaches fall short: existing continual strategies ignore knowledge altogether, while stock neuro-symbolic architectures suffer from catastrophic forgetting. We show that leveraging prior knowledge by combining neuro-symbolic architectures with continual strategies does help avoid catastrophic forgetting, but also that doing so can yield models affected by reasoning shortcuts. These undermine the semantics of the acquired concepts, even when detailed prior knowledge is provided upfront and inference is exact, and in turn continual performance. To overcome these issues, we introduce COOL, a COncept-level cOntinual Learning strategy tailored for neuro-symbolic continual problems that acquires high-quality concepts and remembers them over time. Our experiments on three novel benchmarks highlights how COOL attains sustained high performance on neuro-symbolic continual learning tasks in which other strategies fail.
☆ Diagrammatization: Rationalizing with diagrammatic AI explanations for abductive reasoning on hypotheses
Many visualizations have been developed for explainable AI (XAI), but they often require further reasoning by users to interpret. We argue that XAI should support abductive reasoning - inference to the best explanation - with diagrammatic reasoning to convey hypothesis generation and evaluation. Inspired by Peircean diagrammatic reasoning and the 5-step abduction process, we propose Diagrammatization, an approach to provide diagrammatic, abductive explanations based on domain hypotheses. We implemented DiagramNet for a clinical application to predict diagnoses from heart auscultation, and explain with shape-based murmur diagrams. In modeling studies, we found that DiagramNet not only provides faithful murmur shape explanations, but also has better prediction performance than baseline models. We further demonstrate the usefulness of diagrammatic explanations in a qualitative user study with medical students, showing that clinically-relevant, diagrammatic explanations are preferred over technical saliency map explanations. This work contributes insights into providing domain-conventional abductive explanations for user-centric XAI.
☆ Robust Estimation under the Wasserstein Distance
We study the problem of robust distribution estimation under the Wasserstein metric, a popular discrepancy measure between probability distributions rooted in optimal transport (OT) theory. We introduce a new outlier-robust Wasserstein distance $\mathsf{W}_p^\varepsilon$ which allows for $\varepsilon$ outlier mass to be removed from its input distributions, and show that minimum distance estimation under $\mathsf{W}_p^\varepsilon$ achieves minimax optimal robust estimation risk. Our analysis is rooted in several new results for partial OT, including an approximate triangle inequality, which may be of independent interest. To address computational tractability, we derive a dual formulation for $\mathsf{W}_p^\varepsilon$ that adds a simple penalty term to the classic Kantorovich dual objective. As such, $\mathsf{W}_p^\varepsilon$ can be implemented via an elementary modification to standard, duality-based OT solvers. Our results are extended to sliced OT, where distributions are projected onto low-dimensional subspaces, and applications to homogeneity and independence testing are explored. We illustrate the virtues of our framework via applications to generative modeling with contaminated datasets.
☆ Dual Propagation: Accelerating Contrastive Hebbian Learning with Dyadic Neurons
Activity difference based learning algorithms-such as contrastive Hebbian learning and equilibrium propagation-have been proposed as biologically plausible alternatives to error back-propagation. However, on traditional digital chips these algorithms suffer from having to solve a costly inference problem twice, making these approaches more than two orders of magnitude slower than back-propagation. In the analog realm equilibrium propagation may be promising for fast and energy efficient learning, but states still need to be inferred and stored twice. Inspired by lifted neural networks and compartmental neuron models we propose a simple energy based compartmental neuron model, termed dual propagation, in which each neuron is a dyad with two intrinsic states. At inference time these intrinsic states encode the error/activity duality through their difference and their mean respectively. The advantage of this method is that only a single inference phase is needed and that inference can be solved in layerwise closed-form. Experimentally we show on common computer vision datasets, including Imagenet32x32, that dual propagation performs equivalently to back-propagation both in terms of accuracy and runtime.
☆ Factor Fields: A Unified Framework for Neural Fields and Beyond
We present Factor Fields, a novel framework for modeling and representing signals. Factor Fields decomposes a signal into a product of factors, each of which is represented by a neural or regular field representation operating on a coordinate transformed input signal. We show that this decomposition yields a unified framework that generalizes several recent signal representations including NeRF, PlenOxels, EG3D, Instant-NGP, and TensoRF. Moreover, the framework allows for the creation of powerful new signal representations, such as the Coefficient-Basis Factorization (CoBaFa) which we propose in this paper. As evidenced by our experiments, CoBaFa leads to improvements over previous fast reconstruction methods in terms of the three critical goals in neural signal representation: approximation quality, compactness and efficiency. Experimentally, we demonstrate that our representation achieves better image approximation quality on 2D image regression tasks, higher geometric quality when reconstructing 3D signed distance fields and higher compactness for radiance field reconstruction tasks compared to previous fast reconstruction methods. Besides, our CoBaFa representation enables generalization by sharing the basis across signals during training, enabling generalization tasks such as image regression with sparse observations and few-shot radiance field reconstruction.
comment: 13 pages, 7 figures
☆ Practical Bandits: An Industry Perspective WWW
The bandit paradigm provides a unified modeling framework for problems that require decision-making under uncertainty. Because many business metrics can be viewed as rewards (a.k.a. utilities) that result from actions, bandit algorithms have seen a large and growing interest from industrial applications, such as search, recommendation and advertising. Indeed, with the bandit lens comes the promise of direct optimisation for the metrics we care about. Nevertheless, the road to successfully applying bandits in production is not an easy one. Even when the action space and rewards are well-defined, practitioners still need to make decisions regarding multi-arm or contextual approaches, on- or off-policy setups, delayed or immediate feedback, myopic or long-term optimisation, etc. To make matters worse, industrial platforms typically give rise to large action spaces in which existing approaches tend to break down. The research literature on these topics is broad and vast, but this can overwhelm practitioners, whose primary aim is to solve practical problems, and therefore need to decide on a specific instantiation or approach for each project. This tutorial will take a step towards filling that gap between the theory and practice of bandits. Our goal is to present a unified overview of the field and its existing terminology, concepts and algorithms -- with a focus on problems relevant to industry. We hope our industrial perspective will help future practitioners who wish to leverage the bandit paradigm for their application.
comment: Tutorial held at The Web Conference 2023 (formerly known as WWW) in Austin, Texas (USA), on April 30 - May 4, 2023
☆ Temporal fusion transformer using variational mode decomposition for wind power forecasting
The power output of a wind turbine depends on a variety of factors, including wind speed at different heights, wind direction, temperature and turbine properties. Wind speed and direction, in particular, have complex cycles and fluctuate dramatically, leading to large uncertainties in wind power output. This study uses variational mode decomposition (VMD) to decompose the wind power series and Temporal fusion transformer (TFT) to forecast wind power for the next 1h, 3h and 6h. The experimental results show that VMD outperforms other decomposition algorithms and the TFT model outperforms other decomposition models.
☆ High-precision regressors for particle physics
Monte Carlo simulations of physics processes at particle colliders like the Large Hadron Collider at CERN take up a major fraction of the computational budget. For some simulations, a single data point takes seconds, minutes, or even hours to compute from first principles. Since the necessary number of data points per simulation is on the order of $10^9$ - $10^{12}$, machine learning regressors can be used in place of physics simulators to significantly reduce this computational burden. However, this task requires high-precision regressors that can deliver data with relative errors of less than $1\%$ or even $0.1\%$ over the entire domain of the function. In this paper, we develop optimal training strategies and tune various machine learning regressors to satisfy the high-precision requirement. We leverage symmetry arguments from particle physics to optimize the performance of the regressors. Inspired by ResNets, we design a Deep Neural Network with skip connections that outperform fully connected Deep Neural Networks. We find that at lower dimensions, boosted decision trees far outperform neural networks while at higher dimensions neural networks perform significantly better. We show that these regressors can speed up simulations by a factor of $10^3$ - $10^6$ over the first-principles computations currently used in Monte Carlo simulations. Additionally, using symmetry arguments derived from particle physics, we reduce the number of regressors necessary for each simulation by an order of magnitude. Our work can significantly reduce the training and storage burden of Monte Carlo simulations at current and future collider experiments.
comment: 15 pages, 7 figure and 2 tables
☆ A Theoretical Justification for Image Inpainting using Denoising Diffusion Probabilistic Models
We provide a theoretical justification for sample recovery using diffusion based image inpainting in a linear model setting. While most inpainting algorithms require retraining with each new mask, we prove that diffusion based inpainting generalizes well to unseen masks without retraining. We analyze a recently proposed popular diffusion based inpainting algorithm called RePaint (Lugmayr et al., 2022), and show that it has a bias due to misalignment that hampers sample recovery even in a two-state diffusion process. Motivated by our analysis, we propose a modified RePaint algorithm we call RePaint$^+$ that provably recovers the underlying true sample and enjoys a linear rate of convergence. It achieves this by rectifying the misalignment error present in drift and dispersion of the reverse process. To the best of our knowledge, this is the first linear convergence result for a diffusion based image inpainting algorithm.
comment: 30 pages, 5 figures, 1 Table
☆ Laplacian Change Point Detection for Single and Multi-view Dynamic Graphs
Dynamic graphs are rich data structures that are used to model complex relationships between entities over time. In particular, anomaly detection in temporal graphs is crucial for many real world applications such as intrusion identification in network systems, detection of ecosystem disturbances and detection of epidemic outbreaks. In this paper, we focus on change point detection in dynamic graphs and address three main challenges associated with this problem: i). how to compare graph snapshots across time, ii). how to capture temporal dependencies, and iii). how to combine different views of a temporal graph. To solve the above challenges, we first propose Laplacian Anomaly Detection (LAD) which uses the spectrum of graph Laplacian as the low dimensional embedding of the graph structure at each snapshot. LAD explicitly models short term and long term dependencies by applying two sliding windows. Next, we propose MultiLAD, a simple and effective generalization of LAD to multi-view graphs. MultiLAD provides the first change point detection method for multi-view dynamic graphs. It aggregates the singular values of the normalized graph Laplacian from different views through the scalar power mean operation. Through extensive synthetic experiments, we show that i). LAD and MultiLAD are accurate and outperforms state-of-the-art baselines and their multi-view extensions by a large margin, ii). MultiLAD's advantage over contenders significantly increases when additional views are available, and iii). MultiLAD is highly robust to noise from individual views. In five real world dynamic graphs, we demonstrate that LAD and MultiLAD identify significant events as top anomalies such as the implementation of government COVID-19 interventions which impacted the population mobility in multi-view traffic networks.
comment: 30 pages, 15 figures, extended version of previous paper "Laplacian Change Point Detection for Dynamic Graphs" with novel material. arXiv admin note: substantial text overlap with arXiv:2007.01229
☆ Online Bidding in Repeated Non-Truthful Auctions under Budget and ROI Constraints
Online advertising platforms typically use auction mechanisms to allocate ad placements. Advertisers participate in a series of repeated auctions, and must select bids that will maximize their overall rewards while adhering to certain constraints. We focus on the scenario in which the advertiser has budget and return-on-investment (ROI) constraints. We investigate the problem of budget- and ROI-constrained bidding in repeated non-truthful auctions, such as first-price auctions, and present a best-of-both-worlds framework with no-regret guarantees under both stochastic and adversarial inputs. By utilizing the notion of interval regret, we demonstrate that our framework does not require knowledge of specific parameters of the problem which could be difficult to determine in practice. Our proof techniques can be applied to both the adversarial and stochastic cases with minimal modifications, thereby providing a unified perspective on the two problems. In the adversarial setting, we also show that it is possible to loosen the traditional requirement of having a strictly feasible solution to the offline optimization problem at each round.
☆ Causal Lifting and Link Prediction
Current state-of-the-art causal models for link prediction assume an underlying set of inherent node factors -- an innate characteristic defined at the node's birth -- that governs the causal evolution of links in the graph. In some causal tasks, however, link formation is path-dependent, i.e., the outcome of link interventions depends on existing links. For instance, in the customer-product graph of an online retailer, the effect of an 85-inch TV ad (treatment) likely depends on whether the costumer already has an 85-inch TV. Unfortunately, existing causal methods are impractical in these scenarios. The cascading functional dependencies between links (due to path dependence) are either unidentifiable or require an impractical number of control variables. In order to remedy this shortcoming, this work develops the first causal model capable of dealing with path dependencies in link prediction. It introduces the concept of causal lifting, an invariance in causal models that, when satisfied, allows the identification of causal link prediction queries using limited interventional data. On the estimation side, we show how structural pairwise embeddings -- a type of symmetry-based joint representation of node pairs in a graph -- exhibit lower bias and correctly represent the causal structure of the task, as opposed to existing node embedding methods, e.g., GNNs and matrix factorization. Finally, we validate our theoretical findings on four datasets under three different scenarios for causal link prediction tasks: knowledge base completion, covariance matrix estimation and consumer-product recommendations.
☆ Imitating careful experts to avoid catastrophic events NeurIPS 2022
RL is increasingly being used to control robotic systems that interact closely with humans. This interaction raises the problem of safe RL: how to ensure that a RL-controlled robotic system never, for instance, injures a human. This problem is especially challenging in rich, realistic settings where it is not even possible to clearly write down a reward function which incorporates these outcomes. In these circumstances, perhaps the only viable approach is based on IRL, which infers rewards from human demonstrations. However, IRL is massively underdetermined as many different rewards can lead to the same optimal policies; we show that this makes it difficult to distinguish catastrophic outcomes (such as injuring a human) from merely undesirable outcomes. Our key insight is that humans do display different behaviour when catastrophic outcomes are possible: they become much more careful. We incorporate carefulness signals into IRL, and find that they do indeed allow IRL to disambiguate undesirable from catastrophic outcomes, which is critical to ensuring safety in future real-world human-robot interactions.
comment: 9 pages, 8 figures, accepted to NeurIPS 2022 Workshop on Robot Learning: Trustworthy Robotics
☆ On the Efficacy of Differentially Private Few-shot Image Classification
There has been significant recent progress in training differentially private (DP) models which achieve accuracy that approaches the best non-private models. These DP models are typically pretrained on large public datasets and then fine-tuned on downstream datasets that are (i) relatively large, and (ii) similar in distribution to the pretraining data. However, in many applications including personalization, it is crucial to perform well in the few-shot setting, as obtaining large amounts of labeled data may be problematic; and on images from a wide variety of domains for use in various specialist settings. To understand under which conditions few-shot DP can be effective, we perform an exhaustive set of experiments that reveals how the accuracy and vulnerability to attack of few-shot DP image classification models are affected as the number of shots per class, privacy level, model architecture, dataset, and subset of learnable parameters in the model vary. We show that to achieve DP accuracy on par with non-private models, the shots per class must be increased as the privacy level increases by as much as 32$\times$ for CIFAR-100 at $\epsilon=1$. We also find that few-shot non-private models are highly susceptible to membership inference attacks. DP provides clear mitigation against the attacks, but a small $\epsilon$ is required to effectively prevent them. Finally, we evaluate DP federated learning systems and establish state-of-the-art performance on the challenging FLAIR federated learning benchmark.
☆ Best Possible Q-Learning
Fully decentralized learning, where the global information, i.e., the actions of other agents, is inaccessible, is a fundamental challenge in cooperative multi-agent reinforcement learning. However, the convergence and optimality of most decentralized algorithms are not theoretically guaranteed, since the transition probabilities are non-stationary as all agents are updating policies simultaneously. To tackle this challenge, we propose best possible operator, a novel decentralized operator, and prove that the policies of agents will converge to the optimal joint policy if each agent independently updates its individual state-action value by the operator. Further, to make the update more efficient and practical, we simplify the operator and prove that the convergence and optimality still hold with the simplified one. By instantiating the simplified operator, the derived fully decentralized algorithm, best possible Q-learning (BQL), does not suffer from non-stationarity. Empirically, we show that BQL achieves remarkable improvement over baselines in a variety of cooperative multi-agent tasks.
comment: 14 pages
☆ The Power of Preconditioning in Overparameterized Low-Rank Matrix Sensing
We propose $\textsf{ScaledGD($\lambda$)}$, a preconditioned gradient descent method to tackle the low-rank matrix sensing problem when the true rank is unknown, and when the matrix is possibly ill-conditioned. Using overparametrized factor representations, $\textsf{ScaledGD($\lambda$)}$ starts from a small random initialization, and proceeds by gradient descent with a specific form of damped preconditioning to combat bad curvatures induced by overparameterization and ill-conditioning. At the expense of light computational overhead incurred by preconditioners, $\textsf{ScaledGD($\lambda$)}$ is remarkably robust to ill-conditioning compared to vanilla gradient descent ($\textsf{GD}$) even with overprameterization. Specifically, we show that, under the Gaussian design, $\textsf{ScaledGD($\lambda$)}$ converges to the true low-rank matrix at a constant linear rate after a small number of iterations that scales only logarithmically with respect to the condition number and the problem dimension. This significantly improves over the convergence rate of vanilla $\textsf{GD}$ which suffers from a polynomial dependency on the condition number. Our work provides evidence on the power of preconditioning in accelerating the convergence without hurting generalization in overparameterized learning.
☆ Convolutional Neural Operators
Although very successfully used in machine learning, convolution based neural network architectures -- believed to be inconsistent in function space -- have been largely ignored in the context of learning solution operators of PDEs. Here, we adapt convolutional neural networks to demonstrate that they are indeed able to process functions as inputs and outputs. The resulting architecture, termed as convolutional neural operators (CNOs), is shown to significantly outperform competing models on benchmark experiments, paving the way for the design of an alternative robust and accurate framework for learning operators.
☆ STEP: Learning N:M Structured Sparsity Masks from Scratch with Precondition
Recent innovations on hardware (e.g. Nvidia A100) have motivated learning N:M structured sparsity masks from scratch for fast model inference. However, state-of-the-art learning recipes in this regime (e.g. SR-STE) are proposed for non-adaptive optimizers like momentum SGD, while incurring non-trivial accuracy drop for Adam-trained models like attention-based LLMs. In this paper, we first demonstrate such gap origins from poorly estimated second moment (i.e. variance) in Adam states given by the masked weights. We conjecture that learning N:M masks with Adam should take the critical regime of variance estimation into account. In light of this, we propose STEP, an Adam-aware recipe that learns N:M masks with two phases: first, STEP calculates a reliable variance estimate (precondition phase) and subsequently, the variance remains fixed and is used as a precondition to learn N:M masks (mask-learning phase). STEP automatically identifies the switching point of two phases by dynamically sampling variance changes over the training trajectory and testing the sample concentration. Empirically, we evaluate STEP and other baselines such as ASP and SR-STE on multiple tasks including CIFAR classification, machine translation and LLM fine-tuning (BERT-Base, GPT-2). We show STEP mitigates the accuracy drop of baseline recipes and is robust to aggressive structured sparsity ratios.
☆ Timewarp: Transferable Acceleration of Molecular Dynamics by Learning Time-Coarsened Dynamics
Molecular dynamics (MD) simulation is a widely used technique to simulate molecular systems, most commonly at the all-atom resolution where the equations of motion are integrated with timesteps on the order of femtoseconds ($1\textrm{fs}=10^{-15}\textrm{s}$). MD is often used to compute equilibrium properties, which requires sampling from an equilibrium distribution such as the Boltzmann distribution. However, many important processes, such as binding and folding, occur over timescales of milliseconds or beyond, and cannot be efficiently sampled with conventional MD. Furthermore, new MD simulations need to be performed from scratch for each molecular system studied. We present Timewarp, an enhanced sampling method which uses a normalising flow as a proposal distribution in a Markov chain Monte Carlo method targeting the Boltzmann distribution. The flow is trained offline on MD trajectories and learns to make large steps in time, simulating the molecular dynamics of $10^{5} - 10^{6}\:\textrm{fs}$. Crucially, Timewarp is transferable between molecular systems: once trained, we show that it generalises to unseen small peptides (2-4 amino acids), exploring their metastable states and providing wall-clock acceleration when sampling compared to standard MD. Our method constitutes an important step towards developing general, transferable algorithms for accelerating MD.
☆ Vectorized Scenario Description and Motion Prediction for Scenario-Based Testing
Automated vehicles (AVs) are tested in diverse scenarios, typically specified by parameters such as velocities, distances, or curve radii. To describe scenarios uniformly independent of such parameters, this paper proposes a vectorized scenario description defined by the road geometry and vehicles' trajectories. Data of this form are generated for three scenarios, merged, and used to train the motion prediction model VectorNet, allowing to predict an AV's trajectory for unseen scenarios. Predicting scenario evaluation metrics, VectorNet partially achieves lower errors than regression models that separately process the three scenarios' data. However, for comprehensive generalization, sufficient variance in the training data must be ensured. Thus, contrary to existing methods, our proposed method can merge diverse scenarios' data and exploit spatial and temporal nuances in the vectorized scenario description. As a result, data from specified test scenarios and real-world scenarios can be compared and combined for (predictive) analyses and scenario selection.
comment: 6 pages, 7 figures, 3 tables, submitted to IEEE IV 2023
☆ Deep COVID-19 Forecasting for Multiple States with Data Augmentation
In this work, we propose a deep learning approach to forecasting state-level COVID-19 trends of weekly cumulative death in the United States (US) and incident cases in Germany. This approach includes a transformer model, an ensemble method, and a data augmentation technique for time series. We arrange the inputs of the transformer in such a way that predictions for different states can attend to the trends of the others. To overcome the issue of scarcity of training data for this COVID-19 pandemic, we have developed a novel data augmentation technique to generate useful data for training. More importantly, the generated data can also be used for model validation. As such, it has a two-fold advantage: 1) more actual observations can be used for training, and 2) the model can be validated on data which has distribution closer to the expected situation. Our model has achieved some of the best state-level results on the COVID-19 Forecast Hub for the US and for Germany.
☆ A comparative study of statistical and machine learning models on near-real-time daily emissions prediction
The rapid ascent in carbon dioxide emissions is a major cause of global warming and climate change, which pose a huge threat to human survival and impose far-reaching influence on the global ecosystem. Therefore, it is very necessary to effectively control carbon dioxide emissions by accurately predicting and analyzing the change trend timely, so as to provide a reference for carbon dioxide emissions mitigation measures. This paper is aiming to select a suitable model to predict the near-real-time daily emissions based on univariate daily time-series data from January 1st, 2020 to September 30st, 2022 of all sectors (Power, Industry, Ground Transport, Residential, Domestic Aviation, International Aviation) in China. We proposed six prediction models, which including three statistical models: Grey prediction (GM(1,1)), autoregressive integrated moving average (ARIMA) and seasonal autoregressive integrated moving average with exogenous factors (SARIMAX); three machine learning models: artificial neural network (ANN), random forest (RF) and long short term memory (LSTM). To evaluate the performance of these models, five criteria: Mean Squared Error (MSE), Root Mean Squared Error (RMSE), Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE) and Coefficient of Determination () are imported and discussed in detail. In the results, three machine learning models perform better than that three statistical models, in which LSTM model performs the best on five criteria values for daily emissions prediction with the 3.5179e-04 MSE value, 0.0187 RMSE value, 0.0140 MAE value, 14.8291% MAPE value and 0.9844 value.
☆ UW-CVGAN: UnderWater Image Enhancement with Capsules Vectors Quantization
The degradation in the underwater images is due to wavelength-dependent light attenuation, scattering, and to the diversity of the water types in which they are captured. Deep neural networks take a step in this field, providing autonomous models able to achieve the enhancement of underwater images. We introduce Underwater Capsules Vectors GAN UWCVGAN based on the discrete features quantization paradigm from VQGAN for this task. The proposed UWCVGAN combines an encoding network, which compresses the image into its latent representation, with a decoding network, able to reconstruct the enhancement of the image from the only latent representation. In contrast with VQGAN, UWCVGAN achieves feature quantization by exploiting the clusterization ability of capsule layer, making the model completely trainable and easier to manage. The model obtains enhanced underwater images with high quality and fine details. Moreover, the trained encoder is independent of the decoder giving the possibility to be embedded onto the collector as compressing algorithm to reduce the memory space required for the images, of factor $3\times$. \myUWCVGAN{ }is validated with quantitative and qualitative analysis on benchmark datasets, and we present metrics results compared with the state of the art.
☆ De Novo Molecular Generation via Connection-aware Motif Mining
De novo molecular generation is an essential task for science discovery. Recently, fragment-based deep generative models have attracted much research attention due to their flexibility in generating novel molecules based on existing molecule fragments. However, the motif vocabulary, i.e., the collection of frequent fragments, is usually built upon heuristic rules, which brings difficulties to capturing common substructures from large amounts of molecules. In this work, we propose a new method, MiCaM, to generate molecules based on mined connection-aware motifs. Specifically, it leverages a data-driven algorithm to automatically discover motifs from a molecule library by iteratively merging subgraphs based on their frequency. The obtained motif vocabulary consists of not only molecular motifs (i.e., the frequent fragments), but also their connection information, indicating how the motifs are connected with each other. Based on the mined connection-aware motifs, MiCaM builds a connection-aware generator, which simultaneously picks up motifs and determines how they are connected. We test our method on distribution-learning benchmarks (i.e., generating novel molecules to resemble the distribution of a given training set) and goal-directed benchmarks (i.e., generating molecules with target properties), and achieve significant improvements over previous fragment-based baselines. Furthermore, we demonstrate that our method can effectively mine domain-specific motifs for different tasks.
☆ Mnemosyne: Learning to Train Transformers with Transformers
Training complex machine learning (ML) architectures requires a compute and time consuming process of selecting the right optimizer and tuning its hyper-parameters. A new paradigm of learning optimizers from data has emerged as a better alternative to hand-designed ML optimizers. We propose Mnemosyne optimizer, that uses Performers: implicit low-rank attention Transformers. It can learn to train entire neural network architectures including other Transformers without any task-specific optimizer tuning. We show that Mnemosyne: (a) generalizes better than popular LSTM optimizer, (b) in particular can successfully train Vision Transformers (ViTs) while meta--trained on standard MLPs and (c) can initialize optimizers for faster convergence in Robotics applications. We believe that these results open the possibility of using Transformers to build foundational optimization models that can address the challenges of regular Transformer training. We complement our results with an extensive theoretical analysis of the compact associative memory used by Mnemosyne.
☆ A Survey on Efficient Training of Transformers
Recent advances in Transformers have come with a huge requirement on computing resources, highlighting the importance of developing efficient training techniques to make Transformer training faster, at lower cost, and to higher accuracy by the efficient use of computation and memory resources. This survey provides the first systematic overview of the efficient training of Transformers, covering the recent progress in acceleration arithmetic and hardware, with a focus on the former. We analyze and compare methods that save computation and memory costs for intermediate tensors during training, together with techniques on hardware/algorithm co-design. We finally discuss challenges and promising areas for future research.
comment: A brief review
☆ A general Markov decision process formalism for action-state entropy-regularized reward maximization
Previous work has separately addressed different forms of action, state and action-state entropy regularization, pure exploration and space occupation. These problems have become extremely relevant for regularization, generalization, speeding up learning and providing robust solutions at unprecedented levels. However, solutions of those problems are hectic, ranging from convex and non-convex optimization, and unconstrained optimization to constrained optimization. Here we provide a general dual function formalism that transforms the constrained optimization problem into an unconstrained convex one for any mixture of action and state entropies. The cases with pure action entropy and pure state entropy are understood as limits of the mixture.
☆ Confidence and Dispersity Speak: Characterising Prediction Matrix for Unsupervised Accuracy Estimation
This work aims to assess how well a model performs under distribution shifts without using labels. While recent methods study prediction confidence, this work reports prediction dispersity is another informative cue. Confidence reflects whether the individual prediction is certain; dispersity indicates how the overall predictions are distributed across all categories. Our key insight is that a well-performing model should give predictions with high confidence and high dispersity. That is, we need to consider both properties so as to make more accurate estimates. To this end, we use the nuclear norm that has been shown to be effective in characterizing both properties. Extensive experiments validate the effectiveness of nuclear norm for various models (e.g., ViT and ConvNeXt), different datasets (e.g., ImageNet and CUB-200), and diverse types of distribution shifts (e.g., style shift and reproduction shift). We show that the nuclear norm is more accurate and robust in accuracy estimation than existing methods. Furthermore, we validate the feasibility of other measurements (e.g., mutual information maximization) for characterizing dispersity and confidence. Lastly, we investigate the limitation of the nuclear norm, study its improved variant under severe class imbalance, and discuss potential directions.
comment: This version is not fully edited and will be updated soon
☆ Curriculum Learning for ab initio Deep Learned Refractive Optics
Deep lens optimization has recently emerged as a new paradigm for designing computational imaging systems, however it has been limited to either simple optical systems consisting of a single DOE or metalens, or the fine-tuning of compound lenses from good initial designs. Here we present a deep lens design method based on curriculum learning, which is able to learn optical designs of compound lenses ab initio from randomly initialized surfaces, therefore overcoming the need for a good initial design. We demonstrate this approach with the fully-automatic design of an extended depth-of-field computational camera in a cellphone-style form factor, highly aspherical surfaces, and a short back focal length.
comment: Automatically design computational lenses from scratch with differentiable ray tracing
☆ Uncertainty in Fairness Assessment: Maintaining Stable Conclusions Despite Fluctuations ICML 2023
Several recent works encourage the use of a Bayesian framework when assessing performance and fairness metrics of a classification algorithm in a supervised setting. We propose the Uncertainty Matters (UM) framework that generalizes a Beta-Binomial approach to derive the posterior distribution of any criteria combination, allowing stable performance assessment in a bias-aware setting.We suggest modeling the confusion matrix of each demographic group using a Multinomial distribution updated through a Bayesian procedure. We extend UM to be applicable under the popular K-fold cross-validation procedure. Experiments highlight the benefits of UM over classical evaluation frameworks regarding informativeness and stability.
comment: 25 pages (including references and appendix), 10 figures. Submitted to ICML 2023
☆ MonoFlow: Rethinking Divergence GANs via the Perspective of Differential Equations
The conventional understanding of adversarial training in generative adversarial networks (GANs) is that the discriminator is trained to estimate a divergence, and the generator learns to minimize this divergence. We argue that despite the fact that many variants of GANs were developed following this paradigm, the current theoretical understanding of GANs and their practical algorithms are inconsistent. In this paper, we leverage Wasserstein gradient flows which characterize the evolution of particles in the sample space, to gain theoretical insights and algorithmic inspiration of GANs. We introduce a unified generative modeling framework - MonoFlow: the particle evolution is rescaled via a monotonically increasing mapping of the log density ratio. Under our framework, adversarial training can be viewed as a procedure first obtaining MonoFlow's vector field via training the discriminator and the generator learns to draw the particle flow defined by the corresponding vector field. We also reveal the fundamental difference between variational divergence minimization and adversarial training. This analysis helps us to identify what types of generator loss functions can lead to the successful training of GANs and suggest that GANs may have more loss designs beyond the literature (e.g., non-saturated loss), as long as they realize MonoFlow. Consistent empirical studies are included to validate the effectiveness of our framework.
☆ Fed-GLOSS-DP: Federated, Global Learning using Synthetic Sets with Record Level Differential Privacy
This work proposes Fed-GLOSS-DP, a novel approach to privacy-preserving learning that uses synthetic data to train federated models. In our approach, the server recovers an approximation of the global loss landscape in a local neighborhood based on synthetic samples received from the clients. In contrast to previous, point-wise, gradient-based, linear approximation (such as FedAvg), our formulation enables a type of global optimization that is particularly beneficial in non-IID federated settings. We also present how it rigorously complements record-level differential privacy. Extensive results show that our novel formulation gives rise to considerable improvements in terms of convergence speed and communication costs. We argue that our new approach to federated learning can provide a potential path toward reconciling privacy and accountability by sending differentially private, synthetic data instead of gradient updates. The source code will be released upon publication.
☆ A Survey on Compositional Generalization in Applications
The field of compositional generalization is currently experiencing a renaissance in AI, as novel problem settings and algorithms motivated by various practical applications are being introduced, building on top of the classical compositional generalization problem. This article aims to provide a comprehensive review of top recent developments in multiple real-life applications of the compositional generalization. Specifically, we introduce a taxonomy of common applications and summarize the state-of-the-art for each of those domains. Furthermore, we identify important current trends and provide new perspectives pertaining to the future of this burgeoning field.
☆ Physics Constrained Motion Prediction with Uncertainty Quantification
Predicting the motion of dynamic agents is a critical task for guaranteeing the safety of autonomous systems. A particular challenge is that motion prediction algorithms should obey dynamics constraints and quantify prediction uncertainty as a measure of confidence. We present a physics-constrained approach for motion prediction which uses a surrogate dynamical model to ensure that predicted trajectories are dynamically feasible. We propose a two-step integration consisting of intent and trajectory prediction subject to dynamics constraints. We also construct prediction regions that quantify uncertainty and are tailored for autonomous driving by using conformal prediction, a popular statistical tool. Physics Constrained Motion Prediction achieves a 41% better ADE, 56% better FDE, and 19% better IoU over a baseline in experiments using an autonomous racing dataset.
comment: Submitted to IV 2023
☆ Site-specific Deep Learning Path Loss Models based on the Method of Moments
This paper describes deep learning models based on convolutional neural networks applied to the problem of predicting EM wave propagation over rural terrain. A surface integral equation formulation, solved with the method of moments and accelerated using the Fast Far Field approximation, is used to generate synthetic training data which comprises path loss computed over randomly generated 1D terrain profiles. These are used to train two networks, one based on fractal profiles and one based on profiles generated using a Gaussian process. The models show excellent agreement when applied to test profiles generated using the same statistical process used to create the training data and very good accuracy when applied to real life problems.
comment: EuCAP 2023
☆ Randomized prior wavelet neural operator for uncertainty quantification
In this paper, we propose a novel data-driven operator learning framework referred to as the \textit{Randomized Prior Wavelet Neural Operator} (RP-WNO). The proposed RP-WNO is an extension of the recently proposed wavelet neural operator, which boasts excellent generalizing capabilities but cannot estimate the uncertainty associated with its predictions. RP-WNO, unlike the vanilla WNO, comes with inherent uncertainty quantification module and hence, is expected to be extremely useful for scientists and engineers alike. RP-WNO utilizes randomized prior networks, which can account for prior information and is easier to implement for large, complex deep-learning architectures than its Bayesian counterpart. Four examples have been solved to test the proposed framework, and the results produced advocate favorably for the efficacy of the proposed framework.
☆ Teaching MLOps in Higher Education through Project-Based Learning ICSE
Building and maintaining production-grade ML-enabled components is a complex endeavor that goes beyond the current approach of academic education, focused on the optimization of ML model performance in the lab. In this paper, we present a project-based learning approach to teaching MLOps, focused on the demonstration and experience with emerging practices and tools to automatize the construction of ML-enabled components. We examine the design of a course based on this approach, including laboratory sessions that cover the end-to-end ML component life cycle, from model building to production deployment. Moreover, we report on preliminary results from the first edition of the course. During the present year, an updated version of the same course is being delivered in two independent universities; the related learning outcomes will be evaluated to analyze the effectiveness of project-based learning for this specific subject.
comment: Accepted in 2023 IEEE/ACM 45th International Conference on Software Engineering: Software Engineering Education and Training (ICSE-SEET)
☆ Real-Time Evaluation in Online Continual Learning: A New Paradigm
Current evaluations of Continual Learning (CL) methods typically assume that there is no constraint on training time and computation. This is an unrealistic assumption for any real-world setting, which motivates us to propose: a practical real-time evaluation of continual learning, in which the stream does not wait for the model to complete training before revealing the next data for predictions. To do this, we evaluate current CL methods with respect to their computational costs. We hypothesize that under this new evaluation paradigm, computationally demanding CL approaches may perform poorly on streams with a varying distribution. We conduct extensive experiments on CLOC, a large-scale dataset containing 39 million time-stamped images with geolocation labels. We show that a simple baseline outperforms state-of-the-art CL methods under this evaluation, questioning the applicability of existing methods in realistic settings. In addition, we explore various CL components commonly used in the literature, including memory sampling strategies and regularization approaches. We find that all considered methods fail to be competitive against our simple baseline. This surprisingly suggests that the majority of existing CL literature is tailored to a specific class of streams that is not practical. We hope that the evaluation we provide will be the first step towards a paradigm shift to consider the computational cost in the development of online continual learning methods.
☆ On Suppressing Range of Adaptive Stepsizes of Adam to Improve Generalisation Performance
A number of recent adaptive optimizers improve the generalisation performance of Adam by essentially reducing the variance of adaptive stepsizes to get closer to SGD with momentum. Following the above motivation, we suppress the range of the adaptive stepsizes of Adam by exploiting the layerwise gradient statistics. In particular, at each iteration, we propose to perform three consecutive operations on the second momentum v_t before using it to update a DNN model: (1): down-scaling, (2): epsilon-embedding, and (3): down-translating. The resulting algorithm is referred to as SET-Adam, where SET is a brief notation of the three operations. The down-scaling operation on v_t is performed layerwise by making use of the angles between the layerwise subvectors of v_t and the corresponding all-one subvectors. Extensive experimental results show that SET-Adam outperforms eight adaptive optimizers when training transformers and LSTMs for NLP, and VGG and ResNet for image classification over CIAF10 and CIFAR100 while matching the best performance of the eight adaptive methods when training WGAN-GP models for image generation tasks. Furthermore, SET-Adam produces higher validation accuracies than Adam and AdaBelief for training ResNet18 over ImageNet.
comment: 12 pages. arXiv admin note: substantial text overlap with arXiv:2203.13273
☆ FCB-SwinV2 Transformer for Polyp Segmentation
Polyp segmentation within colonoscopy video frames using deep learning models has the potential to automate the workflow of clinicians. This could help improve the early detection rate and characterization of polyps which could progress to colorectal cancer. Recent state-of-the-art deep learning polyp segmentation models have combined the outputs of Fully Convolutional Network architectures and Transformer Network architectures which work in parallel. In this paper we propose modifications to the current state-of-the-art polyp segmentation model FCBFormer. The transformer architecture of the FCBFormer is replaced with a SwinV2 Transformer-UNET and minor changes to the Fully Convolutional Network architecture are made to create the FCB-SwinV2 Transformer. The performance of the FCB-SwinV2 Transformer is evaluated on the popular colonoscopy segmentation bench-marking datasets Kvasir-SEG and CVC-ClinicDB. Generalizability tests are also conducted. The FCB-SwinV2 Transformer is able to consistently achieve higher mDice scores across all tests conducted and therefore represents new state-of-the-art performance. Issues found with how colonoscopy segmentation model performance is evaluated within literature are also re-ported and discussed. One of the most important issues identified is that when evaluating performance on the CVC-ClinicDB dataset it would be preferable to ensure no data leakage from video sequences occurs during the training/validation/test data partition.
☆ Meta Learning in Decentralized Neural Networks: Towards More General AI AAAI 2023
Meta-learning usually refers to a learning algorithm that learns from other learning algorithms. The problem of uncertainty in the predictions of neural networks shows that the world is only partially predictable and a learned neural network cannot generalize to its ever-changing surrounding environments. Therefore, the question is how a predictive model can represent multiple predictions simultaneously. We aim to provide a fundamental understanding of learning to learn in the contents of Decentralized Neural Networks (Decentralized NNs) and we believe this is one of the most important questions and prerequisites to building an autonomous intelligence machine. To this end, we shall demonstrate several pieces of evidence for tackling the problems above with Meta Learning in Decentralized NNs. In particular, we will present three different approaches to building such a decentralized learning system: (1) learning from many replica neural networks, (2) building the hierarchy of neural networks for different functions, and (3) leveraging different modality experts to learn cross-modal representations.
comment: Accepted for AAAI 2023 workshop
☆ Graph Neural Networks for temporal graphs: State of the art, open challenges, and opportunities
Graph Neural Networks (GNNs) have become the leading paradigm for learning on (static) graph-structured data. However, many real-world systems are dynamic in nature, since the graph and node/edge attributes change over time. In recent years, GNN-based models for temporal graphs have emerged as a promising area of research to extend the capabilities of GNNs. In this work, we provide the first comprehensive overview of the current state-of-the-art of temporal GNN, introducing a rigorous formalization of learning settings and tasks and a novel taxonomy categorizing existing approaches in terms of how the temporal aspect is represented and processed. We conclude the survey with a discussion of the most relevant open challenges for the field, from both research and application perspectives.
☆ Over-parameterised Shallow Neural Networks with Asymmetrical Node Scaling: Global Convergence Guarantees and Feature Learning
We consider the optimisation of large and shallow neural networks via gradient flow, where the output of each hidden node is scaled by some positive parameter. We focus on the case where the node scalings are non-identical, differing from the classical Neural Tangent Kernel (NTK) parameterisation. We prove that, for large neural networks, with high probability, gradient flow converges to a global minimum AND can learn features, unlike in the NTK regime. We also provide experiments on synthetic and real-world datasets illustrating our theoretical results and showing the benefit of such scaling in terms of pruning and transfer learning.
☆ High-Probability Bounds for Stochastic Optimization and Variational Inequalities: the Case of Unbounded Variance
During recent years the interest of optimization and machine learning communities in high-probability convergence of stochastic optimization methods has been growing. One of the main reasons for this is that high-probability complexity bounds are more accurate and less studied than in-expectation ones. However, SOTA high-probability non-asymptotic convergence results are derived under strong assumptions such as the boundedness of the gradient noise variance or of the objective's gradient itself. In this paper, we propose several algorithms with high-probability convergence results under less restrictive assumptions. In particular, we derive new high-probability convergence results under the assumption that the gradient/operator noise has bounded central $\alpha$-th moment for $\alpha \in (1,2]$ in the following setups: (i) smooth non-convex / Polyak-Lojasiewicz / convex / strongly convex / quasi-strongly convex minimization problems, (ii) Lipschitz / star-cocoercive and monotone / quasi-strongly monotone variational inequalities. These results justify the usage of the considered methods for solving problems that do not fit standard functional classes studied in stochastic optimization.
comment: 86 pages
☆ Constrained Online Two-stage Stochastic Optimization: New Algorithms via Adversarial Learning
We consider an online two-stage stochastic optimization with long-term constraints over a finite horizon of $T$ periods. At each period, we take the first-stage action, observe a model parameter realization and then take the second-stage action from a feasible set that depends both on the first-stage decision and the model parameter. We aim to minimize the cumulative objective value while guaranteeing that the long-term average second-stage decision belongs to a set. We propose a general algorithmic framework that derives online algorithms for the online two-stage problem from adversarial learning algorithms. Also, the regret bound of our algorithm cam be reduced to the regret bound of embedded adversarial learning algorithms. Based on our framework, we obtain new results under various settings. When the model parameter at each period is drawn from identical distributions, we derive state-of-art regret bound that improves previous bounds under special cases. Our algorithm is also robust to adversarial corruptions of model parameter realizations. When the model parameters are drawn from unknown non-stationary distributions and we are given prior estimates of the distributions, we develop a new algorithm from our framework with a regret $O(W_T+\sqrt{T})$, where $W_T$ measures the total inaccuracy of the prior estimates.
☆ Unpaired Multi-Domain Causal Representation Learning
The goal of causal representation learning is to find a representation of data that consists of causally related latent variables. We consider a setup where one has access to data from multiple domains that potentially share a causal representation. Crucially, observations in different domains are assumed to be unpaired, that is, we only observe the marginal distribution in each domain but not their joint distribution. In this paper, we give sufficient conditions for identifiability of the joint distribution and the shared causal graph in a linear setup. Identifiability holds if we can uniquely recover the joint distribution and the shared causal representation from the marginal distributions in each domain. We transform our identifiability results into a practical method to recover the shared latent causal graph. Moreover, we study how multiple domains reduce errors in falsely detecting shared causal variables in the finite data setting.
☆ Exposing the CSI: A Systematic Investigation of CSI-based Wi-Fi Sensing Capabilities and Limitations
Thanks to the ubiquitous deployment of Wi-Fi hotspots, channel state information (CSI)-based Wi-Fi sensing can unleash game-changing applications in many fields, such as healthcare, security, and entertainment. However, despite one decade of active research on Wi-Fi sensing, most existing work only considers legacy IEEE 802.11n devices, often in particular and strictly-controlled environments. Worse yet, there is a fundamental lack of understanding of the impact on CSI-based sensing of modern Wi-Fi features, such as 160-MHz bandwidth, multiple-input multiple-output (MIMO) transmissions, and increased spectral resolution in IEEE 802.11ax (Wi-Fi 6). This work aims to shed light on the impact of Wi-Fi 6 features on the sensing performance and to create a benchmark for future research on Wi-Fi sensing. To this end, we perform an extensive CSI data collection campaign involving 3 individuals, 3 environments, and 12 activities, using Wi-Fi 6 signals. An anonymized ground truth obtained through video recording accompanies our 80-GB dataset, which contains almost two hours of CSI data from three collectors. We leverage our dataset to dissect the performance of a state-of-the-art sensing framework across different environments and individuals. Our key findings suggest that (i) MIMO transmissions and higher spectral resolution might be more beneficial than larger bandwidth for sensing applications; (ii) there is a pressing need to standardize research on Wi-Fi sensing because the path towards a truly environment-independent framework is still uncertain. To ease the experiments' replicability and address the current lack of Wi-Fi 6 CSI datasets, we release our 80-GB dataset to the community.
comment: 10 pages, 13 figures, accepted for publication in Proceedings of IEEE PerCom 2023
☆ Speed-Oblivious Online Scheduling: Knowing (Precise) Speeds is not Necessary
We consider online scheduling on unrelated (heterogeneous) machines in a speed-oblivious setting, where an algorithm is unaware of the exact job-dependent processing speeds. We show strong impossibility results for clairvoyant and non-clairvoyant algorithms and overcome them in models inspired by practical settings: (i) we provide competitive learning-augmented algorithms, assuming that (possibly erroneous) predictions on the speeds are given, and (ii) we provide competitive algorithms for the speed-ordered model, where a single global order of machines according to their unknown job-dependent speeds is known. We prove strong theoretical guarantees and evaluate our findings on a representative heterogeneous multi-core processor. These seem to be the first empirical results for algorithms with predictions that are performed in a non-synthetic environment on real hardware.
☆ Predicting Molecule-Target Interaction by Learning Biomedical Network and Molecule Representations
The study of molecule-target interaction is quite important for drug discovery in terms of target identification, pathway study, drug-drug interaction, etc. Most existing methodologies utilize either biomedical network information or molecule structural features to predict potential interaction link. However, the biomedical network information based methods usually suffer from cold start problem, while structure based methods often give limited performance due to the structure/interaction assumption and data quality. To address these issues, we propose a pseudo-siamese Graph Neural Network method, namely MTINet+, which learns both biomedical network topological and molecule structural/chemical information as representations to predict potential interaction of given molecule and target pair. In MTINet+, 1-hop subgraphs of given molecule and target pair are extracted from known interaction of biomedical network as topological information, meanwhile the molecule structural and chemical attributes are processed as molecule information. MTINet+ learns these two types of information as embedding features for predicting the pair link. In the experiments of different molecule-target interaction tasks, MTINet+ significantly outperforms over the state-of-the-art baselines. In addition, in our designed network sparsity experiments , MTINet+ shows strong robustness against different sparse biomedical networks.
comment: 9 pages, 6 figures. arXiv admin note: substantial text overlap with arXiv:2102.01649
☆ Domain Generalization Emerges from Dreaming
Recent studies have proven that DNNs, unlike human vision, tend to exploit texture information rather than shape. Such texture bias is one of the factors for the poor generalization performance of DNNs. We observe that the texture bias negatively affects not only in-domain generalization but also out-of-distribution generalization, i.e., Domain Generalization. Motivated by the observation, we propose a new framework to reduce the texture bias of a model by a novel optimization-based data augmentation, dubbed Stylized Dream. Our framework utilizes adaptive instance normalization (AdaIN) to augment the style of an original image yet preserve the content. We then adopt a regularization loss to predict consistent outputs between Stylized Dream and original images, which encourages the model to learn shape-based representations. Extensive experiments show that the proposed method achieves state-of-the-art performance in out-of-distribution settings on public benchmark datasets: PACS, VLCS, OfficeHome, TerraIncognita, and DomainNet.
comment: 23 pages, 4 figures
☆ A Light-weight CNN Model for Efficient Parkinson's Disease Diagnostics
In recent years, deep learning methods have achieved great success in various fields due to their strong performance in practical applications. In this paper, we present a light-weight neural network for Parkinson's disease diagnostics, in which a series of hand-drawn data are collected to distinguish Parkinson's disease patients from healthy control subjects. The proposed model consists of a convolution neural network (CNN) cascading to long-short-term memory (LSTM) to adapt the characteristics of collected time-series signals. To make full use of their advantages, a multilayered LSTM model is firstly used to enrich features which are then concatenated with raw data and fed into a shallow one-dimensional (1D) CNN model for efficient classification. Experimental results show that the proposed model achieves a high-quality diagnostic result over multiple evaluation metrics with much fewer parameters and operations, outperforming conventional methods such as support vector machine (SVM), random forest (RF), lightgbm (LGB) and CNN-based methods.
☆ Energy Efficiency of Training Neural Network Architectures: An Empirical Study
The evaluation of Deep Learning models has traditionally focused on criteria such as accuracy, F1 score, and related measures. The increasing availability of high computational power environments allows the creation of deeper and more complex models. However, the computations needed to train such models entail a large carbon footprint. In this work, we study the relations between DL model architectures and their environmental impact in terms of energy consumed and CO$_2$ emissions produced during training by means of an empirical study using Deep Convolutional Neural Networks. Concretely, we study: (i) the impact of the architecture and the location where the computations are hosted on the energy consumption and emissions produced; (ii) the trade-off between accuracy and energy efficiency; and (iii) the difference on the method of measurement of the energy consumed using software-based and hardware-based tools.
comment: Accepted in HICSS 2023. For its published version refer to the Proceedings of the 56th Hawaii International Conference on System Sciences; URI https://hdl.handle.net/10125/102727
☆ Resilient Binary Neural Network AAAI 2023
Binary neural networks (BNNs) have received ever-increasing popularity for their great capability of reducing storage burden as well as quickening inference time. However, there is a severe performance drop compared with {real-valued} networks, due to its intrinsic frequent weight oscillation during training. In this paper, we introduce a Resilient Binary Neural Network (ReBNN) to mitigate the frequent oscillation for better BNNs' training. We identify that the weight oscillation mainly stems from the non-parametric scaling factor. To address this issue, we propose to parameterize the scaling factor and introduce a weighted reconstruction loss to build an adaptive training objective. %To the best of our knowledge, it is the first work to solve BNNs based on a dynamically re-weighted loss function. For the first time, we show that the weight oscillation is controlled by the balanced parameter attached to the reconstruction loss, which provides a theoretical foundation to parameterize it in back propagation. Based on this, we learn our ReBNN by {calculating} the {balanced} parameter {based on} its maximum magnitude, which can effectively mitigate the weight oscillation with a resilient training process. Extensive experiments are conducted upon various network models, such as ResNet and Faster-RCNN for computer vision, as well as BERT for natural language processing. The results demonstrate the overwhelming performance of our ReBNN over prior arts. For example, our ReBNN achieves 66.9\% Top-1 accuracy with ResNet-18 backbone on the ImageNet dataset, surpassing existing state-of-the-arts by a significant margin. Our code is open-sourced at https://github.com/SteveTsui/ReBNN.
comment: AAAI 2023 Oral
☆ Deep-Learning Tool for Early Identifying Non-Traumatic Intracranial Hemorrhage Etiology based on CT Scan
Background: To develop an artificial intelligence system that can accurately identify acute non-traumatic intracranial hemorrhage (ICH) etiology based on non-contrast CT (NCCT) scans and investigate whether clinicians can benefit from it in a diagnostic setting. Materials and Methods: The deep learning model was developed with 1868 eligible NCCT scans with non-traumatic ICH collected between January 2011 and April 2018. We tested the model on two independent datasets (TT200 and SD 98) collected after April 2018. The model's diagnostic performance was compared with clinicians's performance. We further designed a simulated study to compare the clinicians's performance with and without the deep learning system augmentation. Results: The proposed deep learning system achieved area under the receiver operating curve of 0.986 (95% CI 0.967-1.000) on aneurysms, 0.952 (0.917-0.987) on hypertensive hemorrhage, 0.950 (0.860-1.000) on arteriovenous malformation (AVM), 0.749 (0.586-0.912) on Moyamoya disease (MMD), 0.837 (0.704-0.969) on cavernous malformation (CM), and 0.839 (0.722-0.959) on other causes in TT200 dataset. Given a 90% specificity level, the sensitivities of our model were 97.1% and 90.9% for aneurysm and AVM diagnosis, respectively. The model also shows an impressive generalizability in an independent dataset SD98. The clinicians achieve significant improvements in the sensitivity, specificity, and accuracy of diagnoses of certain hemorrhage etiologies with proposed system augmentation. Conclusions: The proposed deep learning algorithms can be an effective tool for early identification of hemorrhage etiologies based on NCCT scans. It may also provide more information for clinicians for triage and further imaging examination selection.
☆ Efficient Graph Field Integrators Meet Point Clouds
We present two new classes of algorithms for efficient field integration on graphs encoding point clouds. The first class, SeparatorFactorization(SF), leverages the bounded genus of point cloud mesh graphs, while the second class, RFDiffusion(RFD), uses popular epsilon-nearest-neighbor graph representations for point clouds. Both can be viewed as providing the functionality of Fast Multipole Methods (FMMs), which have had a tremendous impact on efficient integration, but for non-Euclidean spaces. We focus on geometries induced by distributions of walk lengths between points (e.g., shortest-path distance). We provide an extensive theoretical analysis of our algorithms, obtaining new results in structural graph theory as a byproduct. We also perform exhaustive empirical evaluation, including on-surface interpolation for rigid and deformable objects (particularly for mesh-dynamics modeling), Wasserstein distance computations for point clouds, and the Gromov-Wasserstein variant.
☆ An Enhanced V-cycle MgNet Model for Operator Learning in Numerical Partial Differential Equations
This study used a multigrid-based convolutional neural network architecture known as MgNet in operator learning to solve numerical partial differential equations (PDEs). Given the property of smoothing iterations in multigrid methods where low-frequency errors decay slowly, we introduced a low-frequency correction structure for residuals to enhance the standard V-cycle MgNet. The enhanced MgNet model can capture the low-frequency features of solutions considerably better than the standard V-cycle MgNet. The numerical results obtained using some standard operator learning tasks are better than those obtained using many state-of-the-art methods, demonstrating the efficiency of our model.Moreover, numerically, our new model is more robust in case of low- and high-resolution data during training and testing, respectively.
comment: 21 pages, 6 figures, 6 tables
☆ Dynamic Ensemble of Low-fidelity Experts: Mitigating NAS "Cold-Start"
Predictor-based Neural Architecture Search (NAS) employs an architecture performance predictor to improve the sample efficiency. However, predictor-based NAS suffers from the severe ``cold-start'' problem, since a large amount of architecture-performance data is required to get a working predictor. In this paper, we focus on exploiting information in cheaper-to-obtain performance estimations (i.e., low-fidelity information) to mitigate the large data requirements of predictor training. Despite the intuitiveness of this idea, we observe that using inappropriate low-fidelity information even damages the prediction ability and different search spaces have different preferences for low-fidelity information types. To solve the problem and better fuse beneficial information provided by different types of low-fidelity information, we propose a novel dynamic ensemble predictor framework that comprises two steps. In the first step, we train different sub-predictors on different types of available low-fidelity information to extract beneficial knowledge as low-fidelity experts. In the second step, we learn a gating network to dynamically output a set of weighting coefficients conditioned on each input neural architecture, which will be used to combine the predictions of different low-fidelity experts in a weighted sum. The overall predictor is optimized on a small set of actual architecture-performance data to fuse the knowledge from different low-fidelity experts to make the final prediction. We conduct extensive experiments across five search spaces with different architecture encoders under various experimental settings. Our method can easily be incorporated into existing predictor-based NAS frameworks to discover better architectures.
☆ Rethinking Warm-Starts with Predictions: Learning Predictions Close to Sets of Optimal Solutions for Faster $\text{L}$-/$\text{L}^\natural$-Convex Function Minimization
An emerging line of work has shown that machine-learned predictions are useful to warm-start algorithms for discrete optimization problems, such as bipartite matching. Previous studies have shown time complexity bounds proportional to some distance between a prediction and an optimal solution, which we can approximately minimize by learning predictions from past optimal solutions. However, such guarantees may not be meaningful when multiple optimal solutions exist. Indeed, the dual problem of bipartite matching and, more generally, $\text{L}$-/$\text{L}^\natural$-convex function minimization have arbitrarily many optimal solutions, making such prediction-dependent bounds arbitrarily large. To resolve this theoretically critical issue, we present a new warm-start-with-prediction framework for $\text{L}$-/$\text{L}^\natural$-convex function minimization. Our framework offers time complexity bounds proportional to the distance between a prediction and the set of all optimal solutions. The main technical difficulty lies in learning predictions that are provably close to sets of all optimal solutions, for which we present an online-gradient-descent-based method. We thus give the first polynomial-time learnability of predictions that can provably warm-start algorithms regardless of multiple optimal solutions.
☆ LMC: Fast Training of GNNs via Subgraph Sampling with Provable Convergence
The message passing-based graph neural networks (GNNs) have achieved great success in many real-world applications. However, training GNNs on large-scale graphs suffers from the well-known neighbor explosion problem, i.e., the exponentially increasing dependencies of nodes with the number of message passing layers. Subgraph-wise sampling methods -- a promising class of mini-batch training techniques -- discard messages outside the mini-batches in backward passes to avoid the neighbor explosion problem at the expense of gradient estimation accuracy. This poses significant challenges to their convergence analysis and convergence speeds, which seriously limits their reliable real-world applications. To address this challenge, we propose a novel subgraph-wise sampling method with a convergence guarantee, namely Local Message Compensation (LMC). To the best of our knowledge, LMC is the {\it first} subgraph-wise sampling method with provable convergence. The key idea of LMC is to retrieve the discarded messages in backward passes based on a message passing formulation of backward passes. By efficient and effective compensations for the discarded messages in both forward and backward passes, LMC computes accurate mini-batch gradients and thus accelerates convergence. We further show that LMC converges to first-order stationary points of GNNs. Experiments on large-scale benchmark tasks demonstrate that LMC significantly outperforms state-of-the-art subgraph-wise sampling methods in terms of efficiency.
☆ QCM-SGM+: Improved Quantized Compressed Sensing With Score-Based Generative Models for General Sensing Matrices
In realistic compressed sensing (CS) scenarios, the obtained measurements usually have to be quantized to a finite number of bits before transmission and/or storage, thus posing a challenge in recovery, especially for extremely coarse quantization such as 1-bit sign measurements. Recently Meng & Kabashima proposed an efficient quantized compressed sensing algorithm called QCS-SGM using the score-based generative models as an implicit prior. Thanks to the power of score-based generative models in capturing the rich structure of the prior, QCS-SGM achieves remarkably better performances than previous quantized CS methods. However, QCS-SGM is restricted to (approximately) row-orthogonal sensing matrices since otherwise the likelihood score becomes intractable. To address this challenging problem, in this paper we propose an improved version of QCS-SGM, which we call QCS-SGM+, which also works well for general matrices. The key idea is a Bayesian inference perspective of the likelihood score computation, whereby an expectation propagation algorithm is proposed to approximately compute the likelihood score. Experiments on a variety of baseline datasets demonstrate that the proposed QCS-SGM+ outperforms QCS-SGM by a large margin when sensing matrices are far from row-orthogonal.
comment: arXiv admin note: substantial text overlap with arXiv:2211.13006
☆ Conditional expectation for missing data imputation
Missing data is common in datasets retrieved in various areas, such as medicine, sports, and finance. In many cases, to enable proper and reliable analyses of such data, the missing values are often imputed, and it is necessary that the method used has a low root mean square error (RMSE) between the imputed and the true values. In addition, for some critical applications, it is also often a requirement that the logic behind the imputation is explainable, which is especially difficult for complex methods that are for example, based on deep learning. This motivates us to introduce a conditional Distribution based Imputation of Missing Values (DIMV) algorithm. This approach works based on finding the conditional distribution of a feature with missing entries based on the fully observed features. As will be illustrated in the paper, DIMV (i) gives a low RMSE for the imputed values compared to state-of-the-art methods under comparison; (ii) is explainable; (iii) can provide an approximated confidence region for the missing values in a given sample; (iv) works for both small and large scale data; (v) in many scenarios, does not require a huge number of parameters as deep learning approaches and therefore can be used for mobile devices or web browsers; and (vi) is robust to the normally distributed assumption that its theoretical grounds rely on. In addition to DIMV, we also introduce the DPER* algorithm improving the speed of DPER for estimating the mean and covariance matrix from the data, and we confirm the speed-up via experiments.
☆ Energy Efficient Training of SNN using Local Zeroth Order Method
Spiking neural networks are becoming increasingly popular for their low energy requirement in real-world tasks with accuracy comparable to the traditional ANNs. SNN training algorithms face the loss of gradient information and non-differentiability due to the Heaviside function in minimizing the model loss over model parameters. To circumvent the problem surrogate method uses a differentiable approximation of the Heaviside in the backward pass, while the forward pass uses the Heaviside as the spiking function. We propose to use the zeroth order technique at the neuron level to resolve this dichotomy and use it within the automatic differentiation tool. As a result, we establish a theoretical connection between the proposed local zeroth-order technique and the existing surrogate methods and vice-versa. The proposed method naturally lends itself to energy-efficient training of SNNs on GPUs. Experimental results with neuromorphic datasets show that such implementation requires less than 1 percent neurons to be active in the backward pass, resulting in a 100x speed-up in the backward computation time. Our method offers better generalization compared to the state-of-the-art energy-efficient technique while maintaining similar efficiency.
☆ GANalyzer: Analysis and Manipulation of GANs Latent Space for Controllable Face Synthesis
Generative Adversarial Networks (GANs) are capable of synthesizing high-quality facial images. Despite their success, GANs do not provide any information about the relationship between the input vectors and the generated images. Currently, facial GANs are trained on imbalanced datasets, which generate less diverse images. For example, more than 77% of 100K images that we randomly synthesized using the StyleGAN3 are classified as Happy, and only around 3% are Angry. The problem even becomes worse when a mixture of facial attributes is desired: less than 1% of the generated samples are Angry Woman, and only around 2% are Happy Black. To address these problems, this paper proposes a framework, called GANalyzer, for the analysis, and manipulation of the latent space of well-trained GANs. GANalyzer consists of a set of transformation functions designed to manipulate latent vectors for a specific facial attribute such as facial Expression, Age, Gender, and Race. We analyze facial attribute entanglement in the latent space of GANs and apply the proposed transformation for editing the disentangled facial attributes. Our experimental results demonstrate the strength of GANalyzer in editing facial attributes and generating any desired faces. We also create and release a balanced photo-realistic human face dataset. Our code is publicly available on GitHub.
☆ Language Quantized AutoEncoders: Towards Unsupervised Text-Image Alignment
Recent progress in scaling up large language models has shown impressive capabilities in performing few-shot learning across a wide range of text-based tasks. However, a key limitation is that these language models fundamentally lack visual perception - a crucial attribute needed to extend these models to be able to interact with the real world and solve vision tasks, such as in visual-question answering and robotics. Prior works have largely connected image to text through pretraining and/or fine-tuning on curated image-text datasets, which can be a costly and expensive process. In order to resolve this limitation, we propose a simple yet effective approach called Language-Quantized AutoEncoder (LQAE), a modification of VQ-VAE that learns to align text-image data in an unsupervised manner by leveraging pretrained language models (e.g., BERT, RoBERTa). Our main idea is to encode image as sequences of text tokens by directly quantizing image embeddings using a pretrained language codebook. We then apply random masking followed by a BERT model, and have the decoder reconstruct the original image from BERT predicted text token embeddings. By doing so, LQAE learns to represent similar images with similar clusters of text tokens, thereby aligning these two modalities without the use of aligned text-image pairs. This enables few-shot image classification with large language models (e.g., GPT-3) as well as linear classification of images based on BERT text features. To the best of our knowledge, our work is the first work that uses unaligned images for multimodal tasks by leveraging the power of pretrained language models.
comment: arXiv admin note: text overlap with arXiv:2106.13884 by other authors
☆ Quantum Graph Learning: Frontiers and Outlook
Quantum theory has shown its superiority in enhancing machine learning. However, facilitating quantum theory to enhance graph learning is in its infancy. This survey investigates the current advances in quantum graph learning (QGL) from three perspectives, i.e., underlying theories, methods, and prospects. We first look at QGL and discuss the mutualism of quantum theory and graph learning, the specificity of graph-structured data, and the bottleneck of graph learning, respectively. A new taxonomy of QGL is presented, i.e., quantum computing on graphs, quantum graph representation, and quantum circuits for graph neural networks. Pitfall traps are then highlighted and explained. This survey aims to provide a brief but insightful introduction to this emerging field, along with a detailed discussion of frontiers and outlook yet to be investigated.
☆ Neural Common Neighbor with Completion for Link Prediction
Despite its outstanding performance in various graph tasks, vanilla Message Passing Neural Network (MPNN) usually fails in link prediction tasks, as it only uses representations of two individual target nodes and ignores the pairwise relation between them. To capture the pairwise relations, some models add manual features to the input graph and use the output of MPNN to produce pairwise representations. In contrast, others directly use manual features as pairwise representations. Though this simplification avoids applying a GNN to each link individually and thus improves scalability, these models still have much room for performance improvement due to the hand-crafted and unlearnable pairwise features. To upgrade performance while maintaining scalability, we propose Neural Common Neighbor (NCN), which uses learnable pairwise representations. To further boost NCN, we study the unobserved link problem. The incompleteness of the graph is ubiquitous and leads to distribution shifts between the training and test set, loss of common neighbor information, and performance degradation of models. Therefore, we propose two intervention methods: common neighbor completion and target link removal. Combining the two methods with NCN, we propose Neural Common Neighbor with Completion (NCNC). NCN and NCNC outperform recent strong baselines by large margins. NCNC achieves state-of-the-art performance in link prediction tasks.
☆ Synthesizing Physical Character-Scene Interactions
Movement is how people interact with and affect their environment. For realistic character animation, it is necessary to synthesize such interactions between virtual characters and their surroundings. Despite recent progress in character animation using machine learning, most systems focus on controlling an agent's movements in fairly simple and homogeneous environments, with limited interactions with other objects. Furthermore, many previous approaches that synthesize human-scene interactions require significant manual labeling of the training data. In contrast, we present a system that uses adversarial imitation learning and reinforcement learning to train physically-simulated characters that perform scene interaction tasks in a natural and life-like manner. Our method learns scene interaction behaviors from large unstructured motion datasets, without manual annotation of the motion data. These scene interactions are learned using an adversarial discriminator that evaluates the realism of a motion within the context of a scene. The key novelty involves conditioning both the discriminator and the policy networks on scene context. We demonstrate the effectiveness of our approach through three challenging scene interaction tasks: carrying, sitting, and lying down, which require coordination of a character's movements in relation to objects in the environment. Our policies learn to seamlessly transition between different behaviors like idling, walking, and sitting. By randomizing the properties of the objects and their placements during training, our method is able to generalize beyond the objects and scenarios depicted in the training dataset, producing natural character-scene interactions for a wide variety of object shapes and placements. The approach takes physics-based character motion generation a step closer to broad applicability.
☆ Empirical Analysis of the AdaBoost's Error Bound
Understanding the accuracy limits of machine learning algorithms is essential for data scientists to properly measure performance so they can continually improve their models' predictive capabilities. This study empirically verified the error bound of the AdaBoost algorithm for both synthetic and real-world data. The results show that the error bound holds up in practice, demonstrating its efficiency and importance to a variety of applications. The corresponding source code is available at https://github.com/armanbolatov/adaboost_error_bound.
comment: 4 pages, 4 figures
☆ The Contextual Lasso: Sparse Linear Models via Deep Neural Networks
Sparse linear models are a gold standard tool for interpretable machine learning, a field of emerging importance as predictive models permeate decision-making in many domains. Unfortunately, sparse linear models are far less flexible as functions of their input features than black-box models like deep neural networks. With this capability gap in mind, we study a not-uncommon situation where the input features dichotomize into two groups: explanatory features, which we wish to explain the model's predictions, and contextual features, which we wish to determine the model's explanations. This dichotomy leads us to propose the contextual lasso, a new statistical estimator that fits a sparse linear model whose sparsity pattern and coefficients can vary with the contextual features. The fitting process involves learning a nonparametric map, realized via a deep neural network, from contextual feature vector to sparse coefficient vector. To attain sparse coefficients, we train the network with a novel lasso regularizer in the form of a projection layer that maps the network's output onto the space of $\ell_1$-constrained linear models. Extensive experiments on real and synthetic data suggest that the learned models, which remain highly transparent, can be sparser than the regular lasso without sacrificing the predictive power of a standard deep neural network.
☆ Predicting the Silent Majority on Graphs: Knowledge Transferable Graph Neural Network WWW2023
Graphs consisting of vocal nodes ("the vocal minority") and silent nodes ("the silent majority"), namely VS-Graph, are ubiquitous in the real world. The vocal nodes tend to have abundant features and labels. In contrast, silent nodes only have incomplete features and rare labels, e.g., the description and political tendency of politicians (vocal) are abundant while not for ordinary people (silent) on the twitter's social network. Predicting the silent majority remains a crucial yet challenging problem. However, most existing message-passing based GNNs assume that all nodes belong to the same domain, without considering the missing features and distribution-shift between domains, leading to poor ability to deal with VS-Graph. To combat the above challenges, we propose Knowledge Transferable Graph Neural Network (KT-GNN), which models distribution shifts during message passing and representation learning by transferring knowledge from vocal nodes to silent nodes. Specifically, we design the domain-adapted "feature completion and message passing mechanism" for node representation learning while preserving domain difference. And a knowledge transferable classifier based on KL-divergence is followed. Comprehensive experiments on real-world scenarios (i.e., company financial risk assessment and political elections) demonstrate the superior performance of our method. Our source code has been open sourced.
comment: accepted by WWW2023
☆ Reliable Prediction Intervals with Directly Optimized Inductive Conformal Regression for Deep Learning
By generating prediction intervals (PIs) to quantify the uncertainty of each prediction in deep learning regression, the risk of wrong predictions can be effectively controlled. High-quality PIs need to be as narrow as possible, whilst covering a preset proportion of real labels. At present, many approaches to improve the quality of PIs can effectively reduce the width of PIs, but they do not ensure that enough real labels are captured. Inductive Conformal Predictor (ICP) is an algorithm that can generate effective PIs which is theoretically guaranteed to cover a preset proportion of data. However, typically ICP is not directly optimized to yield minimal PI width. However, in this study, we use Directly Optimized Inductive Conformal Regression (DOICR) that takes only the average width of PIs as the loss function and increases the quality of PIs through an optimized scheme under the validity condition that sufficient real labels are captured in the PIs. Benchmark experiments show that DOICR outperforms current state-of-the-art algorithms for regression problems using underlying Deep Neural Network structures for both tabular and image data.
☆ Disentanglement of Latent Representations via Sparse Causal Interventions IJCAI 2023
The process of generating data such as images is controlled by independent and unknown factors of variation. The retrieval of these variables has been studied extensively in the disentanglement, causal representation learning, and independent component analysis fields. Recently, approaches merging these domains together have shown great success. Instead of directly representing the factors of variation, the problem of disentanglement can be seen as finding the interventions on one image that yield a change to a single factor. Following this assumption, we introduce a new method for disentanglement inspired by causal dynamics that combines causality theory with vector-quantized variational autoencoders. Our model considers the quantized vectors as causal variables and links them in a causal graph. It performs causal interventions on the graph and generates atomic transitions affecting a unique factor of variation in the image. We also introduce a new task of action retrieval that consists of finding the action responsible for the transition between two images. We test our method on standard synthetic and real-world disentanglement datasets. We show that it can effectively disentangle the factors of variation and perform precise interventions on high-level semantic attributes of an image without affecting its quality, even with imbalanced data distributions.
comment: 16 pages, 10 pages for the main paper and 6 pages for the supplement, 14 figures, submitted to IJCAI 2023
☆ CLIPood: Generalizing CLIP to Out-of-Distributions
Out-of-distribution (OOD) generalization, where the model needs to handle distribution shifts from training, is a major challenge of machine learning. Recently, contrastive language-image pre-training (CLIP) models have shown impressive zero-shot ability, revealing a promising path toward OOD generalization. However, to boost upon zero-shot performance, further adaptation of CLIP on downstream tasks is indispensable but undesirably degrades OOD generalization ability. In this paper, we aim at generalizing CLIP to out-of-distribution test data on downstream tasks. Beyond the two canonical OOD situations, domain shift and open class, we tackle a more general but difficult in-the-wild setting where both OOD situations may occur on the unseen test data. We propose CLIPood, a simple fine-tuning method that can adapt CLIP models to all OOD situations. To exploit semantic relations between classes from the text modality, CLIPood introduces a new training objective, margin metric softmax (MMS), with class adaptive margins for fine-tuning. Moreover, to incorporate both the pre-trained zero-shot model and the fine-tuned task-adaptive model, CLIPood proposes a new Beta moving average (BMA) to maintain a temporal ensemble according to Beta distribution. Experiments on diverse datasets with different OOD scenarios show that CLIPood consistently outperforms existing generalization techniques.
☆ SimMTM: A Simple Pre-Training Framework for Masked Time-Series Modeling
Time series analysis is widely used in extensive areas. Recently, to reduce labeling expenses and benefit various tasks, self-supervised pre-training has attracted immense interest. One mainstream paradigm is masked modeling, which successfully pre-trains deep models by learning to reconstruct the masked content based on the unmasked part. However, since the semantic information of time series is mainly contained in temporal variations, the standard way of randomly masking a portion of time points will ruin vital temporal variations of time series seriously, making the reconstruction task too difficult to guide representation learning. We thus present SimMTM, a Simple pre-training framework for Masked Time-series Modeling. By relating masked modeling to manifold learning, SimMTM proposes to recover masked time points by the weighted aggregation of multiple neighbors outside the manifold, which eases the reconstruction task by assembling ruined but complementary temporal variations from multiple masked series. SimMTM further learns to uncover the local structure of the manifold helpful for masked modeling. Experimentally, SimMTM achieves state-of-the-art fine-tuning performance in two canonical time series analysis tasks: forecasting and classification, covering both in- and cross-domain settings.
☆ Interventional and Counterfactual Inference with Diffusion Models
We consider the problem of answering observational, interventional, and counterfactual queries in a causally sufficient setting where only observational data and the causal graph are available. Utilizing the recent developments in diffusion models, we introduce diffusion-based causal models (DCM) to learn causal mechanisms, that generate unique latent encodings to allow for direct sampling under interventions as well as abduction for counterfactuals. We utilize DCM to model structural equations, seeing that diffusion models serve as a natural candidate here since they encode each node to a latent representation, a proxy for the exogenous noise, and offer flexible and accurate modeling to provide reliable causal statements and estimates. Our empirical evaluations demonstrate significant improvements over existing state-of-the-art methods for answering causal queries. Our theoretical results provide a methodology for analyzing the counterfactual error for general encoder/decoder models which could be of independent interest.
☆ Algorithm Design for Online Meta-Learning with Task Boundary Detection
Online meta-learning has recently emerged as a marriage between batch meta-learning and online learning, for achieving the capability of quick adaptation on new tasks in a lifelong manner. However, most existing approaches focus on the restrictive setting where the distribution of the online tasks remains fixed with known task boundaries. In this work, we relax these assumptions and propose a novel algorithm for task-agnostic online meta-learning in non-stationary environments. More specifically, we first propose two simple but effective detection mechanisms of task switches and distribution shift based on empirical observations, which serve as a key building block for more elegant online model updates in our algorithm: the task switch detection mechanism allows reusing of the best model available for the current task at hand, and the distribution shift detection mechanism differentiates the meta model update in order to preserve the knowledge for in-distribution tasks and quickly learn the new knowledge for out-of-distribution tasks. In particular, our online meta model updates are based only on the current data, which eliminates the need of storing previous data as required in most existing methods. We further show that a sublinear task-averaged regret can be achieved for our algorithm under mild conditions. Empirical studies on three different benchmarks clearly demonstrate the significant advantage of our algorithm over related baseline approaches.
comment: Submitted for publication
☆ Molecular Geometry-aware Transformer for accurate 3D Atomic System modeling
Molecular dynamic simulations are important in computational physics, chemistry, material, and biology. Machine learning-based methods have shown strong abilities in predicting molecular energy and properties and are much faster than DFT calculations. Molecular energy is at least related to atoms, bonds, bond angles, torsion angles, and nonbonding atom pairs. Previous Transformer models only use atoms as inputs which lack explicit modeling of the aforementioned factors. To alleviate this limitation, we propose Moleformer, a novel Transformer architecture that takes nodes (atoms) and edges (bonds and nonbonding atom pairs) as inputs and models the interactions among them using rotational and translational invariant geometry-aware spatial encoding. Proposed spatial encoding calculates relative position information including distances and angles among nodes and edges. We benchmark Moleformer on OC20 and QM9 datasets, and our model achieves state-of-the-art on the initial state to relaxed energy prediction of OC20 and is very competitive in QM9 on predicting quantum chemical properties compared to other Transformer and Graph Neural Network (GNN) methods which proves the effectiveness of the proposed geometry-aware spatial encoding in Moleformer.
☆ Learning PDE Solution Operator for Continuous Modeling of Time-Series
Learning underlying dynamics from data is important and challenging in many real-world scenarios. Incorporating differential equations (DEs) to design continuous networks has drawn much attention recently, however, most prior works make specific assumptions on the type of DEs, making the model specialized for particular problems. This work presents a partial differential equation (PDE) based framework which improves the dynamics modeling capability. Building upon the recent Fourier neural operator, we propose a neural operator that can handle time continuously without requiring iterative operations or specific grids of temporal discretization. A theoretical result demonstrating its universality is provided. We also uncover an intrinsic property of neural operators that improves data efficiency and model generalization by ensuring stability. Our model achieves superior accuracy in dealing with time-dependent PDEs compared to existing models. Furthermore, several numerical pieces of evidence validate that our method better represents a wide range of dynamics and outperforms state-of-the-art DE-based models in real-time-series applications. Our framework opens up a new way for a continuous representation of neural networks that can be readily adopted for real-world applications.
☆ Convergence of Gradient Descent with Linearly Correlated Noise and Applications to Differentially Private Learning
We study stochastic optimization with linearly correlated noise. Our study is motivated by recent methods for optimization with differential privacy (DP), such as DP-FTRL, which inject noise via matrix factorization mechanisms. We propose an optimization problem that distils key facets of these DP methods and that involves perturbing gradients by linearly correlated noise. We derive improved convergence rates for gradient descent in this framework for convex and non-convex loss functions. Our theoretical analysis is novel and might be of independent interest. We use these convergence rates to develop new, effective matrix factorizations for differentially private optimization, and highlight the benefits of these factorizations theoretically and empirically.
☆ Performance Bounds for Policy-Based Average Reward Reinforcement Learning Algorithms
Many policy-based reinforcement learning (RL) algorithms can be viewed as instantiations of approximate policy iteration (PI), i.e., where policy improvement and policy evaluation are both performed approximately. In applications where the average reward objective is the meaningful performance metric, often discounted reward formulations are used with the discount factor being close to 1, which is equivalent to making the expected horizon very large. However, the corresponding theoretical bounds for error performance scale with the square of the horizon. Thus, even after dividing the total reward by the length of the horizon, the corresponding performance bounds for average reward problems go to infinity. Therefore, an open problem has been to obtain meaningful performance bounds for approximate PI and RL algorithms for the average-reward setting. In this paper, we solve this open problem by obtaining the first non-trivial error bounds for average-reward MDPs which go to zero in the limit where when policy evaluation and policy improvement errors go to zero.
comment: 29 pages
☆ Out of Context: Investigating the Bias and Fairness Concerns of "Artificial Intelligence as a Service"
"AI as a Service" (AIaaS) is a rapidly growing market, offering various plug-and-play AI services and tools. AIaaS enables its customers (users) - who may lack the expertise, data, and/or resources to develop their own systems - to easily build and integrate AI capabilities into their applications. Yet, it is known that AI systems can encapsulate biases and inequalities that can have societal impact. This paper argues that the context-sensitive nature of fairness is often incompatible with AIaaS' 'one-size-fits-all' approach, leading to issues and tensions. Specifically, we review and systematise the AIaaS space by proposing a taxonomy of AI services based on the levels of autonomy afforded to the user. We then critically examine the different categories of AIaaS, outlining how these services can lead to biases or be otherwise harmful in the context of end-user applications. In doing so, we seek to draw research attention to the challenges of this emerging area.
comment: Accepted to CHI '23: ACM Human Factors in Computing, 2023, Hamburg, Germany
☆ Commonsense-Aware Prompting for Controllable Empathetic Dialogue Generation AAAI 2023
Improving the emotional awareness of pre-trained language models is an emerging important problem for dialogue generation tasks. Although prior studies have introduced methods to improve empathetic dialogue generation, few have discussed how to incorporate commonsense knowledge into pre-trained language models for controllable dialogue generation. In this study, we propose a novel framework that improves empathetic dialogue generation using pre-trained language models by 1) incorporating commonsense knowledge through prompt verbalization, and 2) controlling dialogue generation using a strategy-driven future discriminator. We conducted experiments to reveal that both the incorporation of social commonsense knowledge and enforcement of control over generation help to improve generation performance. Finally, we discuss the implications of our study for future research.
comment: Accepted to Workshop on Knowledge Augmented Methods for Natural Language Processing, in conjunction with AAAI 2023
☆ Generalized Uncertainty of Deep Neural Networks: Taxonomy and Applications
Deep neural networks have seen enormous success in various real-world applications. Beyond their predictions as point estimates, increasing attention has been focused on quantifying the uncertainty of their predictions. In this review, we show that the uncertainty of deep neural networks is not only important in a sense of interpretability and transparency, but also crucial in further advancing their performance, particularly in learning systems seeking robustness and efficiency. We will generalize the definition of the uncertainty of deep neural networks to any number or vector that is associated with an input or an input-label pair, and catalog existing methods on ``mining'' such uncertainty from a deep model. We will include those methods from the classic field of uncertainty quantification as well as those methods that are specific to deep neural networks. We then show a wide spectrum of applications of such generalized uncertainty in realistic learning tasks including robust learning such as noisy learning, adversarially robust learning; data-efficient learning such as semi-supervised and weakly-supervised learning; and model-efficient learning such as model compression and knowledge distillation.
☆ Target specific peptide design using latent space approximate trajectory collector
Despite the prevalence and many successes of deep learning applications in de novo molecular design, the problem of peptide generation targeting specific proteins remains unsolved. A main barrier for this is the scarcity of the high-quality training data. To tackle the issue, we propose a novel machine learning based peptide design architecture, called Latent Space Approximate Trajectory Collector (LSATC). It consists of a series of samplers on an optimization trajectory on a highly non-convex energy landscape that approximates the distributions of peptides with desired properties in a latent space. The process involves little human intervention and can be implemented in an end-to-end manner. We demonstrate the model by the design of peptide extensions targeting Beta-catenin, a key nuclear effector protein involved in canonical Wnt signalling. When compared with a random sampler, LSATC can sample peptides with $36\%$ lower binding scores in a $16$ times smaller interquartile range (IQR) and $284\%$ less hydrophobicity with a $1.4$ times smaller IQR. LSATC also largely outperforms other common generative models. Finally, we utilized a clustering algorithm to select 4 peptides from the 100 LSATC designed peptides for experimental validation. The result confirms that all the four peptides extended by LSATC show improved Beta-catenin binding by at least $20.0\%$, and two of the peptides show a $3$ fold increase in binding affinity as compared to the base peptide.
Dataset Distillation Fixes Dataset Reconstruction Attacks
Modern deep learning requires large volumes of data, which could contain sensitive or private information which cannot be leaked. Recent work has shown for homogeneous neural networks a large portion of this training data could be reconstructed with only access to the trained network parameters. While the attack was shown to work empirically, there exists little formal understanding of its effectiveness regime, and ways to defend against it. In this work, we first build a stronger version of the dataset reconstruction attack and show how it can provably recover its entire training set in the infinite width regime. We then empirically study the characteristics of this attack on two-layer networks and reveal that its success heavily depends on deviations from the frozen infinite-width Neural Tangent Kernel limit. More importantly, we formally show for the first time that dataset reconstruction attacks are a variation of dataset distillation. This key theoretical result on the unification of dataset reconstruction and distillation not only sheds more light on the characteristics of the attack but enables us to design defense mechanisms against them via distillation algorithms.
☆ Fast, Differentiable and Sparse Top-k: a Convex Analysis Perspective
The top-k operator returns a k-sparse vector, where the non-zero values correspond to the k largest values of the input. Unfortunately, because it is a discontinuous function, it is difficult to incorporate in neural networks trained end-to-end with backpropagation. Recent works have considered differentiable relaxations, based either on regularization or perturbation techniques. However, to date, no approach is fully differentiable and sparse. In this paper, we propose new differentiable and sparse top-k operators. We view the top-k operator as a linear program over the permutahedron, the convex hull of permutations. We then introduce a p-norm regularization term to smooth out the operator, and show that its computation can be reduced to isotonic optimization. Our framework is significantly more general than the existing one and allows for example to express top-k operators that select values in magnitude. On the algorithmic side, in addition to pool adjacent violator (PAV) algorithms, we propose a new GPU/TPU-friendly Dykstra algorithm to solve isotonic optimization problems. We successfully use our operators to prune weights in neural networks, to fine-tune vision transformers, and as a router in sparse mixture of experts.
comment: 23 pages
☆ A Convolutional-based Model for Early Prediction of Alzheimer's based on the Dementia Stage in the MRI Brain Images
Alzheimer's disease is a degenerative brain disease. Being the primary cause of Dementia in adults and progressively destroys brain memory. Though Alzheimer's disease does not have a cure currently, diagnosing it at an earlier stage will help reduce the severity of the disease. Thus, early diagnosis of Alzheimer's could help to reduce or stop the disease from progressing. In this paper, we proposed a deep convolutional neural network-based model for learning model using to determine the stage of Dementia in adults based on the Magnetic Resonance Imaging (MRI) images to detect the early onset of Alzheimer's.
comment: Short paper, Under Review in FLAIRS-36
☆ Neural Insights for Digital Marketing Content Design
In digital marketing, experimenting with new website content is one of the key levers to improve customer engagement. However, creating successful marketing content is a manual and time-consuming process that lacks clear guiding principles. This paper seeks to close the loop between content creation and online experimentation by offering marketers AI-driven actionable insights based on historical data to improve their creative process. We present a neural-network-based system that scores and extracts insights from a marketing content design, namely, a multimodal neural network predicts the attractiveness of marketing contents, and a post-hoc attribution method generates actionable insights for marketers to improve their content in specific marketing locations. Our insights not only point out the advantages and drawbacks of a given current content, but also provide design recommendations based on historical data. We show that our scoring model and insights work well both quantitatively and qualitatively.
☆ Hyperbolic Contrastive Learning
Learning good image representations that are beneficial to downstream tasks is a challenging task in computer vision. As such, a wide variety of self-supervised learning approaches have been proposed. Among them, contrastive learning has shown competitive performance on several benchmark datasets. The embeddings of contrastive learning are arranged on a hypersphere that results in using the inner (dot) product as a distance measurement in Euclidean space. However, the underlying structure of many scientific fields like social networks, brain imaging, and computer graphics data exhibit highly non-Euclidean latent geometry. We propose a novel contrastive learning framework to learn semantic relationships in the hyperbolic space. Hyperbolic space is a continuous version of trees that naturally owns the ability to model hierarchical structures and is thus beneficial for efficient contrastive representation learning. We also extend the proposed Hyperbolic Contrastive Learning (HCL) to the supervised domain and studied the adversarial robustness of HCL. The comprehensive experiments show that our proposed method achieves better results on self-supervised pretraining, supervised classification, and higher robust accuracy than baseline methods.
♻ ☆ "Why did the Model Fail?": Attributing Model Performance Changes to Distribution Shifts
Performance of machine learning models may differ between training and deployment for many reasons. For instance, model performance can change between environments due to changes in data quality, observing a different population than the one in training, or changes in the relationship between labels and features. These changes result in distribution shifts across environments. Attributing model performance changes to specific shifts is critical for identifying sources of model failures, and for taking mitigating actions that ensure robust models. In this work, we introduce the problem of attributing performance differences between environments to distribution shifts in the underlying data generating mechanisms. We formulate the problem as a cooperative game where the players are distributions. We define the value of a set of distributions to be the change in model performance when only this set of distributions has changed between environments, and derive an importance weighting method for computing the value of an arbitrary set of distributions. The contribution of each distribution to the total performance change is then quantified as its Shapley value. We demonstrate the correctness and utility of our method on synthetic, semi-synthetic, and real-world case studies, showing its effectiveness in attributing performance changes to a wide range of distribution shifts.
♻ ☆ Analysis of Knowledge Transfer in Kernel Regime CIKM 2022
Knowledge transfer is shown to be a very successful technique for training neural classifiers: together with the ground truth data, it uses the "privileged information" (PI) obtained by a "teacher" network to train a "student" network. It has been observed that classifiers learn much faster and more reliably via knowledge transfer. However, there has been little or no theoretical analysis of this phenomenon. To bridge this gap, we propose to approach the problem of knowledge transfer by regularizing the fit between the teacher and the student with PI provided by the teacher. Using tools from dynamical systems theory, we show that when the student is an extremely wide two layer network, we can analyze it in the kernel regime and show that it is able to interpolate between PI and the given data. This characterization sheds new light on the relation between the training error and capacity of the student relative to the teacher. Another contribution of the paper is a quantitative statement on the convergence of student network. We prove that the teacher reduces the number of required iterations for a student to learn, and consequently improves the generalization power of the student. We give corresponding experimental analysis that validates the theoretical results and yield additional insights.
comment: The work is published by CIKM 2022
♻ ☆ On the Feasibility of Machine Learning Augmented Magnetic Resonance for Point-of-Care Identification of Disease
Early detection of many life-threatening diseases (e.g., prostate and breast cancer) within at-risk population can improve clinical outcomes and reduce cost of care. While numerous disease-specific "screening" tests that are closer to Point-of-Care (POC) are in use for this task, their low specificity results in unnecessary biopsies, leading to avoidable patient trauma and wasteful healthcare spending. On the other hand, despite the high accuracy of Magnetic Resonance (MR) imaging in disease diagnosis, it is not used as a POC disease identification tool because of poor accessibility. The root cause of poor accessibility of MR stems from the requirement to reconstruct high-fidelity images, as it necessitates a lengthy and complex process of acquiring large quantities of high-quality k-space measurements. In this study we explore the feasibility of an ML-augmented MR pipeline that directly infers the disease sidestepping the image reconstruction process. We hypothesise that the disease classification task can be solved using a very small tailored subset of k-space data, compared to image reconstruction. Towards that end, we propose a method that performs two tasks: 1) identifies a subset of the k-space that maximizes disease identification accuracy, and 2) infers the disease directly using the identified k-space subset, bypassing the image reconstruction step. We validate our hypothesis by measuring the performance of the proposed system across multiple diseases and anatomies. We show that comparable performance to image-based classifiers, trained on images reconstructed with full k-space data, can be achieved using small quantities of data: 8% of the data for detecting multiple abnormalities in prostate and brain scans, and 5% of the data for knee abnormalities. To better understand the proposed approach and instigate future research, we provide an extensive analysis and release code.
♻ ☆ A Closer Look at Few-shot Classification Again
Few-shot classification consists of a training phase where a model is learned on a relatively large dataset and an adaptation phase where the learned model is adapted to previously-unseen tasks with limited labeled samples. In this paper, we empirically prove that the training algorithm and the adaptation algorithm can be completely disentangled, which allows algorithm analysis and design to be done individually for each phase. Our meta-analysis for each phase reveals several interesting insights that may help better understand key aspects of few-shot classification and connections with other fields such as visual representation learning and transfer learning. We hope the insights and research challenges revealed in this paper can inspire future work in related directions.
♻ ☆ Box$^2$EL: Concept and Role Box Embeddings for the Description Logic EL++
Representation learning in the form of semantic embeddings has been successfully applied to a variety of tasks in natural language processing and knowledge graphs. Recently, there has been growing interest in developing similar methods for learning embeddings of entire ontologies. We propose Box$^2$EL, a novel method for representation learning of ontologies in the Description Logic EL++, which represents both concepts and roles as boxes (i.e. axis-aligned hyperrectangles), such that the logical structure of the ontology is preserved. We theoretically prove the soundness of our model and conduct an extensive empirical evaluation, in which we achieve state-of-the-art results in subsumption prediction, link prediction, and deductive reasoning. As part of our evaluation, we introduce a novel benchmark for evaluating EL++ embedding models on predicting subsumptions involving both atomic and complex concepts.
comment: Corrected the GitHub URL and updated baselines
♻ ☆ An Instrumental Variable Approach to Confounded Off-Policy Evaluation
Off-policy evaluation (OPE) is a method for estimating the return of a target policy using some pre-collected observational data generated by a potentially different behavior policy. In some cases, there may be unmeasured variables that can confound the action-reward or action-next-state relationships, rendering many existing OPE approaches ineffective. This paper develops an instrumental variable (IV)-based method for consistent OPE in confounded Markov decision processes (MDPs). Similar to single-stage decision making, we show that IV enables us to correctly identify the target policy's value in infinite horizon settings as well. Furthermore, we propose an efficient and robust value estimator and illustrate its effectiveness through extensive simulations and analysis of real data from a world-leading short-video platform.
♻ ☆ Live in the Moment: Learning Dynamics Model Adapted to Evolving Policy
Model-based reinforcement learning (RL) often achieves higher sample efficiency in practice than model-free RL by learning a dynamics model to generate samples for policy learning. Previous works learn a dynamics model that fits under the empirical state-action visitation distribution for all historical policies, i.e., the sample replay buffer. However, in this paper, we observe that fitting the dynamics model under the distribution for \emph{all historical policies} does not necessarily benefit model prediction for the \emph{current policy} since the policy in use is constantly evolving over time. The evolving policy during training will cause state-action visitation distribution shifts. We theoretically analyze how this distribution shift over historical policies affects the model learning and model rollouts. We then propose a novel dynamics model learning method, named \textit{Policy-adapted Dynamics Model Learning (PDML)}. PDML dynamically adjusts the historical policy mixture distribution to ensure the learned model can continually adapt to the state-action visitation distribution of the evolving policy. Experiments on a range of continuous control environments in MuJoCo show that PDML achieves significant improvement in sample efficiency and higher asymptotic performance combined with the state-of-the-art model-based RL methods.
comment: 16 pages, 5 figures
♻ ☆ Prediction-Powered Inference
We introduce prediction-powered inference $\unicode{x2013}$ a framework for performing valid statistical inference when an experimental data set is supplemented with predictions from a machine-learning system. Our framework yields provably valid conclusions without making any assumptions on the machine-learning algorithm that supplies the predictions. Higher accuracy of the predictions translates to smaller confidence intervals, permitting more powerful inference. Prediction-powered inference yields simple algorithms for computing valid confidence intervals for statistical objects such as means, quantiles, and linear and logistic regression coefficients. We demonstrate the benefits of prediction-powered inference with data sets from proteomics, genomics, electronic voting, remote sensing, census analysis, and ecology.
comment: Code is available at https://github.com/aangelopoulos/prediction-powered-inference
♻ ☆ Training Normalizing Flows with the Precision-Recall Divergence
Generative models can have distinct mode of failures like mode dropping and low quality samples, which cannot be captured by a single scalar metric. To address this, recent works propose evaluating generative models using precision and recall, where precision measures quality of samples and recall measures the coverage of the target distribution. Although a variety of discrepancy measures between the target and estimated distribution are used to train generative models, it is unclear what precision-recall trade-offs are achieved by various choices of the discrepancy measures. In this paper, we show that achieving a specified precision-recall trade-off corresponds to minimising -divergences from a family we call the {\em PR-divergences }. Conversely, any -divergence can be written as a linear combination of PR-divergences and therefore correspond to minimising a weighted precision-recall trade-off. Further, we propose a novel generative model that is able to train a normalizing flow to minimise any -divergence, and in particular, achieve a given precision-recall trade-off.
♻ ☆ Large-scale Stochastic Optimization of NDCG Surrogates for Deep Learning with Provable Convergence ICML2022
NDCG, namely Normalized Discounted Cumulative Gain, is a widely used ranking metric in information retrieval and machine learning. However, efficient and provable stochastic methods for maximizing NDCG are still lacking, especially for deep models. In this paper, we propose a principled approach to optimize NDCG and its top-$K$ variant. First, we formulate a novel compositional optimization problem for optimizing the NDCG surrogate, and a novel bilevel compositional optimization problem for optimizing the top-$K$ NDCG surrogate. Then, we develop efficient stochastic algorithms with provable convergence guarantees for the non-convex objectives. Different from existing NDCG optimization methods, the per-iteration complexity of our algorithms scales with the mini-batch size instead of the number of total items. To improve the effectiveness for deep learning, we further propose practical strategies by using initial warm-up and stop gradient operator. Experimental results on multiple datasets demonstrate that our methods outperform prior ranking approaches in terms of NDCG. To the best of our knowledge, this is the first time that stochastic algorithms are proposed to optimize NDCG with a provable convergence guarantee. Our proposed methods are implemented in the LibAUC library at https://libauc.org/.
comment: 32 pages, 12 figures; Accepted by ICML2022
♻ ☆ Combinatorial Inference on the Optimal Assortment in Multinomial Logit Models
Assortment optimization has received active explorations in the past few decades due to its practical importance. Despite the extensive literature dealing with optimization algorithms and latent score estimation, uncertainty quantification for the optimal assortment still needs to be explored and is of great practical significance. Instead of estimating and recovering the complete optimal offer set, decision-makers may only be interested in testing whether a given property holds true for the optimal assortment, such as whether they should include several products of interest in the optimal set, or how many categories of products the optimal set should include. This paper proposes a novel inferential framework for testing such properties. We consider the widely adopted multinomial logit (MNL) model, where we assume that each customer will purchase an item within the offered products with a probability proportional to the underlying preference score associated with the product. We reduce inferring a general optimal assortment property to quantifying the uncertainty associated with the sign change point detection of the marginal revenue gaps. We show the asymptotic normality of the marginal revenue gap estimator, and construct a maximum statistic via the gap estimators to detect the sign change point. By approximating the distribution of the maximum statistic with multiplier bootstrap techniques, we propose a valid testing procedure. We also conduct numerical experiments to assess the performance of our method.
♻ ☆ BridgeTower: Building Bridges Between Encoders in Vision-Language Representation Learning AAAI 2023
Vision-Language (VL) models with the Two-Tower architecture have dominated visual-language representation learning in recent years. Current VL models either use lightweight uni-modal encoders and learn to extract, align and fuse both modalities simultaneously in a deep cross-modal encoder, or feed the last-layer uni-modal representations from the deep pre-trained uni-modal encoders into the top cross-modal encoder. Both approaches potentially restrict vision-language representation learning and limit model performance. In this paper, we propose BridgeTower, which introduces multiple bridge layers that build a connection between the top layers of uni-modal encoders and each layer of the cross-modal encoder. This enables effective bottom-up cross-modal alignment and fusion between visual and textual representations of different semantic levels of pre-trained uni-modal encoders in the cross-modal encoder. Pre-trained with only 4M images, BridgeTower achieves state-of-the-art performance on various downstream vision-language tasks. In particular, on the VQAv2 test-std set, BridgeTower achieves an accuracy of 78.73%, outperforming the previous state-of-the-art model METER by 1.09% with the same pre-training data and almost negligible additional parameters and computational costs. Notably, when further scaling the model, BridgeTower achieves an accuracy of 81.15%, surpassing models that are pre-trained on orders-of-magnitude larger datasets. Code and checkpoints are available at https://github.com/microsoft/BridgeTower.
comment: Accepted by AAAI 2023, Oral
♻ ☆ Epistemic Neural Networks
Intelligence relies on an agent's knowledge of what it does not know. This capability can be assessed based on the quality of joint predictions of labels across multiple inputs. In principle, ensemble-based approaches produce effective joint predictions, but the computational costs of training large ensembles can become prohibitive. We introduce the epinet: an architecture that can supplement any conventional neural network, including large pretrained models, and can be trained with modest incremental computation to estimate uncertainty. With an epinet, conventional neural networks outperform very large ensembles, consisting of hundreds or more particles, with orders of magnitude less computation. The epinet does not fit the traditional framework of Bayesian neural networks. To accommodate development of approaches beyond BNNs, such as the epinet, we introduce the epistemic neural network (ENN) as an interface for models that produce joint predictions.
♻ ☆ Optimal Stopping via Randomized Neural Networks
This paper presents new machine learning approaches to approximate the solutions of optimal stopping problems. The key idea of these methods is to use neural networks, where the parameters of the hidden layers are generated randomly and only the last layer is trained, in order to approximate the continuation value. Our approaches are applicable to high dimensional problems where the existing approaches become increasingly impractical. In addition, since our approaches can be optimized using simple linear regression, they are easy to implement and theoretical guarantees are provided. Our randomized reinforcement learning approach and randomized recurrent neural network approach outperform the state-of-the-art and other relevant machine learning approaches in Markovian and non-Markovian examples, respectively. In particular, we test our approaches on Black-Scholes, Heston, rough Heston and fractional Brownian motion. Moreover, we show that they can also be used to efficiently compute Greeks of American options.
♻ ☆ Do Kernel and Neural Embeddings Help in Training and Generalization?
Recent results on optimization and generalization properties of neural networks showed that in a simple two-layer network, the alignment of the labels to the eigenvectors of the corresponding Gram matrix determines the convergence of the optimization during training. Such analyses also provide upper bounds on the generalization error. We experimentally investigate the implications of these results to deeper networks via embeddings. We regard the layers preceding the final hidden layer as producing different representations of the input data which are then fed to the two-layer model. We show that these representations improve both optimization and generalization. In particular, we investigate three kernel representations when fed to the final hidden layer: the Gaussian kernel and its approximation by random Fourier features, kernels designed to imitate representations produced by neural networks and finally an optimal kernel designed to align the data with target labels. The approximated representations induced by these kernels are fed to the neural network and the optimization and generalization properties of the final model are evaluated and compared.
comment: This work is published by Neural Processing Letters
♻ ☆ On the Effective Number of Linear Regions in Shallow Univariate ReLU Networks: Convergence Guarantees and Implicit Bias
We study the dynamics and implicit bias of gradient flow (GF) on univariate ReLU neural networks with a single hidden layer in a binary classification setting. We show that when the labels are determined by the sign of a target network with $r$ neurons, with high probability over the initialization of the network and the sampling of the dataset, GF converges in direction (suitably defined) to a network achieving perfect training accuracy and having at most $\mathcal{O}(r)$ linear regions, implying a generalization bound. Unlike many other results in the literature, under an additional assumption on the distribution of the data, our result holds even for mild over-parameterization, where the width is $\tilde{\mathcal{O}}(r)$ and independent of the sample size.
♻ ☆ Survey of Knowledge Distillation in Federated Edge Learning
The increasing demand for intelligent services and privacy protection of mobile and Internet of Things (IoT) devices motivates the wide application of Federated Edge Learning (FEL), in which devices collaboratively train on-device Machine Learning (ML) models without sharing their private data. Limited by device hardware, diverse user behaviors and network infrastructure, the algorithm design of FEL faces challenges related to resources, personalization and network environments. Fortunately, Knowledge Distillation (KD) has been leveraged as an important technique to tackle the above challenges in FEL. In this paper, we investigate the works that KD applies to FEL, discuss the limitations and open problems of existing KD-based FEL approaches, and provide guidance for their real deployment.
comment: 13 pages, 1 figure, 2 tables
♻ ☆ A Light Recipe to Train Robust Vision Transformers
In this paper, we ask whether Vision Transformers (ViTs) can serve as an underlying architecture for improving the adversarial robustness of machine learning models against evasion attacks. While earlier works have focused on improving Convolutional Neural Networks, we show that also ViTs are highly suitable for adversarial training to achieve competitive performance. We achieve this objective using a custom adversarial training recipe, discovered using rigorous ablation studies on a subset of the ImageNet dataset. The canonical training recipe for ViTs recommends strong data augmentation, in part to compensate for the lack of vision inductive bias of attention modules, when compared to convolutions. We show that this recipe achieves suboptimal performance when used for adversarial training. In contrast, we find that omitting all heavy data augmentation, and adding some additional bag-of-tricks ($\varepsilon$-warmup and larger weight decay), significantly boosts the performance of robust ViTs. We show that our recipe generalizes to different classes of ViT architectures and large-scale models on full ImageNet-1k. Additionally, investigating the reasons for the robustness of our models, we show that it is easier to generate strong attacks during training when using our recipe and that this leads to better robustness at test time. Finally, we further study one consequence of adversarial training by proposing a way to quantify the semantic nature of adversarial perturbations and highlight its correlation with the robustness of the model. Overall, we recommend that the community should avoid translating the canonical training recipes in ViTs to robust training and rethink common training choices in the context of adversarial training.
comment: Camera-ready version for SaTML 2023, code available at https://github.com/dedeswim/vits-robustness-torch
♻ ☆ Optimization-Based Separations for Neural Networks
Depth separation results propose a possible theoretical explanation for the benefits of deep neural networks over shallower architectures, establishing that the former possess superior approximation capabilities. However, there are no known results in which the deeper architecture leverages this advantage into a provable optimization guarantee. We prove that when the data are generated by a distribution with radial symmetry which satisfies some mild assumptions, gradient descent can efficiently learn ball indicator functions using a depth 2 neural network with two layers of sigmoidal activations, and where the hidden layer is held fixed throughout training. By building on and refining existing techniques for approximation lower bounds of neural networks with a single layer of non-linearities, we show that there are $d$-dimensional radial distributions on the data such that ball indicators cannot be learned efficiently by any algorithm to accuracy better than $\Omega(d^{-4})$, nor by a standard gradient descent implementation to accuracy better than a constant. These results establish what is to the best of our knowledge, the first optimization-based separations where the approximation benefits of the stronger architecture provably manifest in practice. Our proof technique introduces new tools and ideas that may be of independent interest in the theoretical study of both the approximation and optimization of neural networks.
♻ ☆ Planckian Jitter: countering the color-crippling effects of color jitter on self-supervised training ICLR 2023
Several recent works on self-supervised learning are trained by mapping different augmentations of the same image to the same feature representation. The data augmentations used are of crucial importance to the quality of learned feature representations. In this paper, we analyze how the color jitter traditionally used in data augmentation negatively impacts the quality of the color features in learned feature representations. To address this problem, we propose a more realistic, physics-based color data augmentation - which we call Planckian Jitter - that creates realistic variations in chromaticity and produces a model robust to illumination changes that can be commonly observed in real life, while maintaining the ability to discriminate image content based on color information. Experiments confirm that such a representation is complementary to the representations learned with the currently-used color jitter augmentation and that a simple concatenation leads to significant performance gains on a wide range of downstream datasets. In addition, we present a color sensitivity analysis that documents the impact of different training methods on model neurons and shows that the performance of the learned features is robust with respect to illuminant variations.
comment: Accepted at Eleventh International Conference on Learning Representations (ICLR 2023)
♻ ☆ An Advantage Using Feature Selection with a Quantum Annealer
Feature selection is a technique in statistical prediction modeling that identifies features in a record with a strong statistical connection to the target variable. Excluding features with a weak statistical connection to the target variable in training not only drops the dimension of the data, which decreases the time complexity of the algorithm, it also decreases noise within the data which assists in avoiding overfitting. In all, feature selection assists in training a robust statistical model that performs well and is stable. Given the lack of scalability in classical computation, current techniques only consider the predictive power of the feature and not redundancy between the features themselves. Recent advancements in feature selection that leverages quantum annealing (QA) gives a scalable technique that aims to maximize the predictive power of the features while minimizing redundancy. As a consequence, it is expected that this algorithm would assist in the bias/variance trade-off yielding better features for training a statistical model. This paper tests this intuition against classical methods by utilizing open-source data sets and evaluate the efficacy of each trained statistical model well-known prediction algorithms. The numerical results display an advantage utilizing the features selected from the algorithm that leveraged QA.
♻ ☆ Neural Design for Genetic Perturbation Experiments
The problem of how to genetically modify cells in order to maximize a certain cellular phenotype has taken center stage in drug development over the last few years (with, for example, genetically edited CAR-T, CAR-NK, and CAR-NKT cells entering cancer clinical trials). Exhausting the search space for all possible genetic edits (perturbations) or combinations thereof is infeasible due to cost and experimental limitations. This work provides a theoretically sound framework for iteratively exploring the space of perturbations in pooled batches in order to maximize a target phenotype under an experimental budget. Inspired by this application domain, we study the problem of batch query bandit optimization and introduce the Optimistic Arm Elimination ($\mathrm{OAE}$) principle designed to find an almost optimal arm under different functional relationships between the queries (arms) and the outputs (rewards). We analyze the convergence properties of $\mathrm{OAE}$ by relating it to the Eluder dimension of the algorithm's function class and validate that $\mathrm{OAE}$ outperforms other strategies in finding optimal actions in experiments on simulated problems, public datasets well-studied in bandit contexts, and in genetic perturbation datasets when the regression model is a deep neural network. OAE also outperforms the benchmark algorithms in 3 of 4 datasets in the GeneDisco experimental planning challenge.
comment: 22 pages main, 15 pages appendix
♻ ☆ Safe Optimization of an Industrial Refrigeration Process Using an Adaptive and Explorative Framework
Many industrial applications rely on real-time optimization to improve key performance indicators. In the case of unknown process characteristics, real-time optimization becomes challenging, particularly for the satisfaction of safety constraints. In this paper, we demonstrate the application of an adaptive and explorative real-time optimization framework to an industrial refrigeration process, where we learn the process characteristics through changes in process control targets and through exploration to satisfy safety constraints. We quantify the uncertainty in unknown compressor characteristics of the refrigeration plant by using Gaussian processes and incorporate this uncertainty into the objective function of the real-time optimization problem as a weighted cost term. We adaptively control the weight of this term to drive exploration. The results of our simulation experiments indicate the proposed approach can help to increase the energy efficiency of the considered refrigeration process, closely approximating the performance of a solution that has complete information about the compressor performance characteristics.
comment: Under review for IFAC WC 2023. arXiv admin note: substantial text overlap with arXiv:2211.05495
♻ ☆ Unravelling the Performance of Physics-informed Graph Neural Networks for Dynamical Systems NeurIPS 2022
Recently, graph neural networks have been gaining a lot of attention to simulate dynamical systems due to their inductive nature leading to zero-shot generalizability. Similarly, physics-informed inductive biases in deep-learning frameworks have been shown to give superior performance in learning the dynamics of physical systems. There is a growing volume of literature that attempts to combine these two approaches. Here, we evaluate the performance of thirteen different graph neural networks, namely, Hamiltonian and Lagrangian graph neural networks, graph neural ODE, and their variants with explicit constraints and different architectures. We briefly explain the theoretical formulation highlighting the similarities and differences in the inductive biases and graph architecture of these systems. We evaluate these models on spring, pendulum, gravitational, and 3D deformable solid systems to compare the performance in terms of rollout error, conserved quantities such as energy and momentum, and generalizability to unseen system sizes. Our study demonstrates that GNNs with additional inductive biases, such as explicit constraints and decoupling of kinetic and potential energies, exhibit significantly enhanced performance. Further, all the physics-informed GNNs exhibit zero-shot generalizability to system sizes an order of magnitude larger than the training system, thus providing a promising route to simulate large-scale realistic systems.
comment: Accepted at 36th Conference on Neural Information Processing Systems (NeurIPS 2022)
♻ ☆ Modelling the long-term fairness dynamics of data-driven targeted help on job seekers
The use of data-driven decision support by public agencies is becoming more widespread and already influences the allocation of public resources. This raises ethical concerns, as it has adversely affected minorities and historically discriminated groups. In this paper, we use an approach that combines statistics and data-driven approaches with dynamical modeling to assess long-term fairness effects of labor market interventions. Specifically, we develop and use a model to investigate the impact of decisions caused by a public employment authority that selectively supports job-seekers through targeted help. The selection of who receives what help is based on a data-driven intervention model that estimates an individual's chances of finding a job in a timely manner and rests upon data that describes a population in which skills relevant to the labor market are unevenly distributed between two groups (e.g., males and females). The intervention model has incomplete access to the individual's actual skills and can augment this with knowledge of the individual's group affiliation, thus using a protected attribute to increase predictive accuracy. We assess this intervention model's dynamics -- especially fairness-related issues and trade-offs between different fairness goals -- over time and compare it to an intervention model that does not use group affiliation as a predictive feature. We conclude that in order to quantify the trade-off correctly and to assess the long-term fairness effects of such a system in the real-world, careful modeling of the surrounding labor market is indispensable.
♻ ☆ Adversarial Networks and Machine Learning for File Classification
Correctly identifying the type of file under examination is a critical part of a forensic investigation. The file type alone suggests the embedded content, such as a picture, video, manuscript, spreadsheet, etc. In cases where a system owner might desire to keep their files inaccessible or file type concealed, we propose using an adversarially-trained machine learning neural network to determine a file's true type even if the extension or file header is obfuscated to complicate its discovery. Our semi-supervised generative adversarial network (SGAN) achieved 97.6% accuracy in classifying files across 11 different types. We also compared our network against a traditional standalone neural network and three other machine learning algorithms. The adversarially-trained network proved to be the most precise file classifier especially in scenarios with few supervised samples available. Our implementation of a file classifier using an SGAN is implemented on GitHub (https://ksaintg.github.io/SGAN-File-Classier).
♻ ☆ Dynamic Recognition of Speakers for Consent Management by Contrastive Embedding Replay
Voice assistants overhear conversations and a consent management mechanism is required. Consent management can be implemented using speaker recognition. Users that do not give consent enrol their voice and all their further recordings are discarded. Building speaker recognition-based consent management is challenging as dynamic registration, removal, and re-registration of speakers must be efficiently handled. This work proposes a consent management system addressing the aforementioned challenges. A contrastive based training is applied to learn the underlying speaker equivariance inductive bias. The contrastive features for buckets of speakers are trained a few steps into each iteration and act as replay buffers. These features are progressively selected using a multi-strided random sampler for classification. Moreover, new methods for dynamic registration using a portion of old utterances, removal, and re-registration of speakers are proposed. The results verify memory efficiency and dynamic capabilities of the proposed methods and outperform the existing approach from the literature.
comment: This work has been submitted to the IEEE for possible publication. Copyright may be transferred without notice, after which this version may no longer be accessible. The current version includes 36 pages, 8 figures, and 3 tables
♻ ☆ Is it worth it? Comparing six deep and classical methods for unsupervised anomaly detection in time series
Detecting anomalies in time series data is important in a variety of fields, including system monitoring, healthcare, and cybersecurity. While the abundance of available methods makes it difficult to choose the most appropriate method for a given application, each method has its strengths in detecting certain types of anomalies. In this study, we compare six unsupervised anomaly detection methods of varying complexity to determine whether more complex methods generally perform better and if certain methods are better suited to certain types of anomalies. We evaluated the methods using the UCR anomaly archive, a recent benchmark dataset for anomaly detection. We analyzed the results on a dataset and anomaly type level after adjusting the necessary hyperparameters for each method. Additionally, we assessed the ability of each method to incorporate prior knowledge about anomalies and examined the differences between point-wise and sequence-wise features. Our experiments show that classical machine learning methods generally outperform deep learning methods across a range of anomaly types.
comment: 17 Pages, The repository to reproduce the results is available at https://gitlab.com/dlr-dw/is-it-worth-it-benchmark
♻ ☆ Augmentation Component Analysis: Modeling Similarity via the Augmentation Overlaps ICLR 2023
Self-supervised learning aims to learn a embedding space where semantically similar samples are close. Contrastive learning methods pull views of samples together and push different samples away, which utilizes semantic invariance of augmentation but ignores the relationship between samples. To better exploit the power of augmentation, we observe that semantically similar samples are more likely to have similar augmented views. Therefore, we can take the augmented views as a special description of a sample. In this paper, we model such a description as the augmentation distribution and we call it augmentation feature. The similarity in augmentation feature reflects how much the views of two samples overlap and is related to their semantical similarity. Without computational burdens to explicitly estimate values of the augmentation feature, we propose Augmentation Component Analysis (ACA) with a contrastive-like loss to learn principal components and an on-the-fly projection loss to embed data. ACA equals an efficient dimension reduction by PCA and extracts low-dimensional embeddings, theoretically preserving the similarity of augmentation distribution between samples. Empirical results show our method can achieve competitive results against various traditional contrastive learning methods on different benchmarks.
comment: Accept to ICLR 2023
♻ ☆ Symbolic Physics Learner: Discovering governing equations via Monte Carlo tree search
Nonlinear dynamics is ubiquitous in nature and commonly seen in various science and engineering disciplines. Distilling analytical expressions that govern nonlinear dynamics from limited data remains vital but challenging. To tackle this fundamental issue, we propose a novel Symbolic Physics Learner (SPL) machine to discover the mathematical structure of nonlinear dynamics. The key concept is to interpret mathematical operations and system state variables by computational rules and symbols, establish symbolic reasoning of mathematical formulas via expression trees, and employ a Monte Carlo tree search (MCTS) agent to explore optimal expression trees based on measurement data. The MCTS agent obtains an optimistic selection policy through the traversal of expression trees, featuring the one that maps to the arithmetic expression of underlying physics. Salient features of the proposed framework include search flexibility and enforcement of parsimony for discovered equations. The efficacy and superiority of the SPL machine are demonstrated by numerical examples, compared with state-of-the-art baselines.
comment: 22 pages
♻ ☆ A survey, review, and future trends of skin lesion segmentation and classification
The Computer-aided Diagnosis or Detection (CAD) approach for skin lesion analysis is an emerging field of research that has the potential to alleviate the burden and cost of skin cancer screening. Researchers have recently indicated increasing interest in developing such CAD systems, with the intention of providing a user-friendly tool to dermatologists to reduce the challenges encountered or associated with manual inspection. This article aims to provide a comprehensive literature survey and review of a total of 594 publications (356 for skin lesion segmentation and 238 for skin lesion classification) published between 2011 and 2022. These articles are analyzed and summarized in a number of different ways to contribute vital information regarding the methods for the development of CAD systems. These ways include relevant and essential definitions and theories, input data (dataset utilization, preprocessing, augmentations, and fixing imbalance problems), method configuration (techniques, architectures, module frameworks, and losses), training tactics (hyperparameter settings), and evaluation criteria. We intend to investigate a variety of performance-enhancing approaches, including ensemble and post-processing. We also discuss these dimensions to reveal their current trends based on utilization frequencies. In addition, we highlight the primary difficulties associated with evaluating skin lesion segmentation and classification systems using minimal datasets, as well as the potential solutions to these difficulties. Findings, recommendations, and trends are disclosed to inform future research on developing an automated and robust CAD system for skin lesion analysis.
comment: This manuscript has been accepted to be published in Computers in Biology and Medicine and has a total of 106 pages (single column and double spacing), 13 figures, and 11 tables
♻ ☆ An adversarial feature learning strategy for debiasing neural networks
Simplicity bias is the concerning tendency of deep networks to over-depend on simple, weakly predictive features, to the exclusion of stronger, more complex features. This causes biased, incorrect model predictions in many real-world applications, exacerbated by incomplete training data containing spurious feature-label correlations. We propose a direct, interventional method for addressing simplicity bias in DNNs, which we call the feature sieve. We aim to automatically identify and suppress easily-computable spurious features in lower layers of the network, thereby allowing the higher network levels to extract and utilize richer, more meaningful representations. We provide concrete evidence of this differential suppression & enhancement of relevant features on both controlled datasets and real-world images, and report substantial gains on many real-world debiasing benchmarks (11.4% relative gain on Imagenet-A; 3.2% on BAR, etc). Crucially, we outperform many baselines that incorporate knowledge about known spurious or biased attributes, despite our method not using any such information. We believe that our feature sieve work opens up exciting new research directions in automated adversarial feature extraction & representation learning for deep networks.
♻ ☆ Error estimates for physics informed neural networks approximating the Navier-Stokes equations
We prove rigorous bounds on the errors resulting from the approximation of the incompressible Navier-Stokes equations with (extended) physics informed neural networks. We show that the underlying PDE residual can be made arbitrarily small for tanh neural networks with two hidden layers. Moreover, the total error can be estimated in terms of the training error, network size and number of quadrature points. The theory is illustrated with numerical experiments.
♻ ☆ Improved Analysis of Score-based Generative Modeling: User-Friendly Bounds under Minimal Smoothness Assumptions
We give an improved theoretical analysis of score-based generative modeling. Under a score estimate with small $L^2$ error (averaged across timesteps), we provide efficient convergence guarantees for any data distribution with second-order moment, by either employing early stopping or assuming smoothness condition on the score function of the data distribution. Our result does not rely on any log-concavity or functional inequality assumption and has a logarithmic dependence on the smoothness. In particular, we show that under only a finite second moment condition, approximating the following in reverse KL divergence in $\epsilon$-accuracy can be done in $\tilde O\left(\frac{d \log (1/\delta)}{\epsilon}\right)$ steps: 1) the variance-$\delta$ Gaussian perturbation of any data distribution; 2) data distributions with $1/\delta$-smooth score functions. Our analysis also provides a quantitative comparison between different discrete approximations and may guide the choice of discretization points in practice.
comment: 36 pages
♻ ☆ A Scalable, Interpretable, Verifiable & Differentiable Logic Gate Convolutional Neural Network Architecture From Truth Tables
We propose $\mathcal{T}$ruth $\mathcal{T}$able net ($\mathcal{TT}$net), a novel Convolutional Neural Network (CNN) architecture that addresses, by design, the open challenges of interpretability, formal verification, and logic gate conversion. $\mathcal{TT}$net is built using CNNs' filters that are equivalent to tractable truth tables and that we call Learning Truth Table (LTT) blocks. The dual form of LTT blocks allows the truth tables to be easily trained with gradient descent and makes these CNNs easy to interpret, verify and infer. Specifically, $\mathcal{TT}$net is a deep CNN model that can be automatically represented, after post-training transformation, as a sum of Boolean decision trees, or as a sum of Disjunctive/Conjunctive Normal Form (DNF/CNF) formulas, or as a compact Boolean logic circuit. We demonstrate the effectiveness and scalability of $\mathcal{TT}$net on multiple datasets, showing comparable interpretability to decision trees, fast complete/sound formal verification, and scalable logic gate representation, all compared to state-of-the-art methods. We believe this work represents a step towards making CNNs more transparent and trustworthy for real-world critical applications.
♻ ☆ Encouraging Intra-Class Diversity Through a Reverse Contrastive Loss for Better Single-Source Domain Generalization
Traditional deep learning algorithms often fail to generalize when they are tested outside of the domain of the training data. The issue can be mitigated by using unlabeled data from the target domain at training time, but because data distributions can change dynamically in real-life applications once a learned model is deployed, it is critical to create networks robust to unknown and unforeseen domain shifts. In this paper we focus on one of the reasons behind the inability of neural networks to be so: deep networks focus only on the most obvious, potentially spurious, clues to make their predictions and are blind to useful but slightly less efficient or more complex patterns. This behaviour has been identified and several methods partially addressed the issue. To investigate their effectiveness and limits, we first design a publicly available MNIST-based benchmark to precisely measure the ability of an algorithm to find the ''hidden'' patterns. Then, we evaluate state-of-the-art algorithms through our benchmark and show that the issue is largely unsolved. Finally, we propose a partially reversed contrastive loss to encourage intra-class diversity and find less strongly correlated patterns, whose efficiency is demonstrated by our experiments.
♻ ☆ LogLG: Weakly Supervised Log Anomaly Detection via Log-Event Graph Construction
Fully supervised log anomaly detection methods suffer the heavy burden of annotating massive unlabeled log data. Recently, many semi-supervised methods have been proposed to reduce annotation costs with the help of parsed templates. However, these methods consider each keyword independently, which disregards the correlation between keywords and the contextual relationships among log sequences. In this paper, we propose a novel weakly supervised log anomaly detection framework, named LogLG, to explore the semantic connections among keywords from sequences. Specifically, we design an end-to-end iterative process, where the keywords of unlabeled logs are first extracted to construct a log-event graph. Then, we build a subgraph annotator to generate pseudo labels for unlabeled log sequences. To ameliorate the annotation quality, we adopt a self-supervised task to pre-train a subgraph annotator. After that, a detection model is trained with the generated pseudo labels. Conditioned on the classification results, we re-extract the keywords from the log sequences and update the log-event graph for the next iteration. Experiments on five benchmarks validate the effectiveness of LogLG for detecting anomalies on unlabeled log data and demonstrate that LogLG, as the state-of-the-art weakly supervised method, achieves significant performance improvements compared to existing methods.
comment: 12 pages
♻ ☆ Towards Label-Efficient Incremental Learning: A Survey
The current dominant paradigm when building a machine learning model is to iterate over a dataset over and over until convergence. Such an approach is non-incremental, as it assumes access to all images of all categories at once. However, for many applications, non-incremental learning is unrealistic. To that end, researchers study incremental learning, where a learner is required to adapt to an incoming stream of data with a varying distribution while preventing forgetting of past knowledge. Significant progress has been made, however, the vast majority of works focus on the fully supervised setting, making these algorithms label-hungry thus limiting their real-life deployment. To that end, in this paper, we make the first attempt to survey recently growing interest in label-efficient incremental learning. We identify three subdivisions, namely semi-, few-shot- and self-supervised learning to reduce labeling efforts. Finally, we identify novel directions that can further enhance label-efficiency and improve incremental learning scalability. Project website: https://github.com/kilickaya/label-efficient-il.
♻ ☆ BRAIxDet: Learning to Detect Malignant Breast Lesion with Incomplete Annotations
Methods to detect malignant lesions from screening mammograms are usually trained with fully annotated datasets, where images are labelled with the localisation and classification of cancerous lesions. However, real-world screening mammogram datasets commonly have a subset that is fully annotated and another subset that is weakly annotated with just the global classification (i.e., without lesion localisation). Given the large size of such datasets, researchers usually face a dilemma with the weakly annotated subset: to not use it or to fully annotate it. The first option will reduce detection accuracy because it does not use the whole dataset, and the second option is too expensive given that the annotation needs to be done by expert radiologists. In this paper, we propose a middle-ground solution for the dilemma, which is to formulate the training as a weakly- and semi-supervised learning problem that we refer to as malignant breast lesion detection with incomplete annotations. To address this problem, our new method comprises two stages, namely: 1) pre-training a multi-view mammogram classifier with weak supervision from the whole dataset, and 2) extending the trained classifier to become a multi-view detector that is trained with semi-supervised student-teacher learning, where the training set contains fully and weakly-annotated mammograms. We provide extensive detection results on two real-world screening mammogram datasets containing incomplete annotations, and show that our proposed approach achieves state-of-the-art results in the detection of malignant breast lesions with incomplete annotations.
comment: Under Review
♻ ☆ Provably Doubly Accelerated Federated Learning: The First Theoretically Successful Combination of Local Training and Communication Compression
In federated learning, a large number of users are involved in a global learning task, in a collaborative way. They alternate local computations and two-way communication with a distant orchestrating server. Communication, which can be slow and costly, is the main bottleneck in this setting. To reduce the communication load and therefore accelerate distributed gradient descent, two strategies are popular: 1) communicate less frequently; that is, perform several iterations of local computations between the communication rounds; and 2) communicate compressed information instead of full-dimensional vectors. We propose the first algorithm for distributed optimization and federated learning, which harnesses these two strategies jointly and converges linearly to an exact solution in the strongly convex setting, with a doubly accelerated rate: our algorithm benefits from the two acceleration mechanisms provided by local training and compression, namely a better dependency on the condition number of the functions and on the dimension of the model, respectively.
♻ ☆ From Traditional Adaptive Data Caching to Adaptive Context Caching: A Survey
Context data is in demand more than ever with the rapid increase in the development of many context-aware Internet of Things applications. Research in context and context-awareness is being conducted to broaden its applicability in light of many practical and technical challenges. One of the challenges is improving performance when responding to large number of context queries. Context Management Platforms that infer and deliver context to applications measure this problem using Quality of Service (QoS) parameters. Although caching is a proven way to improve QoS, transiency of context and features such as variability, heterogeneity of context queries pose an additional real-time cost management problem. This paper presents a critical survey of state-of-the-art in adaptive data caching with the objective of developing a body of knowledge in cost- and performance-efficient adaptive caching strategies. We comprehensively survey a large number of research publications and evaluate, compare, and contrast different techniques, policies, approaches, and schemes in adaptive caching. Our critical analysis is motivated by the focus on adaptively caching context as a core research problem. A formal definition for adaptive context caching is then proposed, followed by identified features and requirements of a well-designed, objective optimal adaptive context caching strategy.
♻ ☆ Distillation Policy Optimization
On-policy algorithms are supposed to be stable, however, sample-intensive yet. Off-policy algorithms utilizing past experiences are deemed to be sample-efficient, nevertheless, unstable in general. Can we design an algorithm that can employ the off-policy data, while exploit the stable learning by sailing along the course of the on-policy walkway? In this paper, we present an actor-critic learning framework that borrows the distributional perspective of interest to evaluate, and cross-breeds two sources of the data for policy improvement, which enables fast learning and can be applied to a wide class of algorithms. In its backbone, the variance reduction mechanisms, such as unified advantage estimator (UAE), that extends generalized advantage estimator (GAE) to be applicable on any state-dependent baseline, and a learned baseline, that is competent to stabilize the policy gradient, are firstly put forward to not merely be a bridge to the action-value function but also distill the advantageous learning signal. Lastly, it is empirically shown that our method improves sample efficiency and interpolates different levels well. Being of an organic whole, its mixture places more inspiration to the algorithm design.
♻ ☆ Truthful Self-Play ICLR 2023
We present a general framework for evolutionary learning to emergent unbiased state representation without any supervision. Evolutionary frameworks such as self-play converge to bad local optima in case of multi-agent reinforcement learning in non-cooperative partially observable environments with communication due to information asymmetry. Our proposed framework is a simple modification of self-play inspired by mechanism design, also known as {\em reverse game theory}, to elicit truthful signals and make the agents cooperative. The key idea is to add imaginary rewards using the peer prediction method, i.e., a mechanism for evaluating the validity of information exchanged between agents in a decentralized environment. Numerical experiments with predator prey, traffic junction and StarCraft tasks demonstrate that the state-of-the-art performance of our framework.
comment: Accepted for publication at ICLR 2023
♻ ☆ Homotopy-based training of NeuralODEs for accurate dynamics discovery ICML2023
Conceptually, Neural Ordinary Differential Equations (NeuralODEs) pose an attractive way to extract dynamical laws from time series data, as they are natural extensions of the traditional differential equation-based modeling paradigm of the physical sciences. In practice, NeuralODEs display long training times and suboptimal results, especially for longer duration data where they may fail to fit the data altogether. While methods have been proposed to stabilize NeuralODE training, many of these involve placing a strong constraint on the functional form the trained NeuralODE can take that the actual underlying governing equation does not guarantee satisfaction. In this work, we present a novel NeuralODE training algorithm that leverages tools from the chaos and mathematical optimization communities - synchronization and homotopy optimization - for a breakthrough in tackling the NeuralODE training obstacle. We demonstrate architectural changes are unnecessary for effective NeuralODE training. Compared to the conventional training methods, our algorithm achieves drastically lower loss values without any changes to the model architectures. Experiments on both simulated and real systems with complex temporal behaviors demonstrate NeuralODEs trained with our algorithm are able to accurately capture true long term behaviors and correctly extrapolate into the future.
comment: 19 pages, 17 figures, submitted to ICML2023
♻ ☆ Information-theoretic limitations of data-based price discrimination
This paper studies third-degree price discrimination (3PD) based on a random sample of valuation and covariate data, where the covariate is continuous, and the distribution of the data is unknown to the seller. The main results of this paper are twofold. The first set of results is pricing strategy independent and reveals the fundamental information-theoretic limitation of any data-based pricing strategy in revenue generation for two cases: 3PD and uniform pricing. The second set of results proposes the $K$-markets empirical revenue maximization (ERM) strategy and shows that the $K$-markets ERM and the uniform ERM strategies achieve the optimal rate of convergence in revenue to that generated by their respective true-distribution 3PD and uniform pricing optima. Our theoretical and numerical results suggest that the uniform (i.e., $1$-market) ERM strategy generates a larger revenue than the $K$-markets ERM strategy when the sample size is small enough, and vice versa.
comment: In the new version, we have (1) added a simulation and empirical study and (2) fixed some minor issues and improved the clarity
♻ ☆ Stochastic Optimization for Counterfactual Explanations
Explainable AI offers insights into what factors drive a certain prediction of a black-box AI system. One popular interpreting approach is through counterfactual explanations, which go beyond why a system arrives at a certain decision to further provide suggestions on what a user can do to alter the original outcome. A counterfactual example must possess plenty of desiderata. These constraints exist at trade-offs between one and another presenting radical challenges to existing works. We here propose a stochastic learning-based framework that effectively balances the counterfactual trade-offs. The framework consists of a generation and a feature selection module with complementary roles: the former aims to model the distribution of valid counterfactuals whereas the latter serves to enforce additional constraints in a way that allows for differentiable training and amortized optimization. We demonstrate the effectiveness of our method in generating actionable and plausible counterfactuals that are more diverse than the existing methods and particularly more efficient than closest baselines.
♻ ☆ A Survey of Methods, Challenges and Perspectives in Causality
The Causality field aims to find systematic methods for uncovering cause-effect relationships. Such methods can find applications in many research fields, justifying a great interest in this domain. Machine Learning models have shown success in a large variety of tasks by extracting correlation patterns from high-dimensional data but still struggle when generalizing out of their initial distribution. As causal engines aim to learn mechanisms that are independent from a data distribution, combining Machine Learning with Causality has the potential to bring benefits to the two fields. In our work, we motivate this assumption and provide applications. We first perform an extensive overview of the theories and methods for Causality from different perspectives. We then provide a deeper look at the connections between Causality and Machine Learning and describe the challenges met by the two domains. We show the early attempts to bring the fields together and the possible perspectives for the future. We finish by providing a large variety of applications for techniques from Causality.
comment: 40 pages, 37 pages for the main paper and 3 pages for the supplement, 10 figures, submitted to ACM Computing Surveys; enabled hyperlinks in new version
♻ ☆ The KFIoU Loss for Rotated Object Detection ICLR 2023
Differing from the well-developed horizontal object detection area whereby the computing-friendly IoU based loss is readily adopted and well fits with the detection metrics. In contrast, rotation detectors often involve a more complicated loss based on SkewIoU which is unfriendly to gradient-based training. In this paper, we propose an effective approximate SkewIoU loss based on Gaussian modeing and Kalman filter, which mainly consists of two items. The first term is a scale-insensitive center point loss, which is used to quickly get the center points between bounding boxes closer to assist the second term. In the distance-independent second term, Kalman filter is adopted to inherently mimic the mechanism of SkewIoU by its definition, and show its alignment with the SkewIoU loss at trend-level within a certain distance (i.e. within 9 pixels). This is in contrast to recent Gaussian modeling based rotation detectors e.g. GWD loss and KLD loss that involve a human-specified distribution distance metric which require additional hyperparameter tuning that vary across datasets and detectors. The resulting new loss called KFIoU loss is easier to implement and works better compared with exact SkewIoU loss, thanks to its full differentiability and ability to handle the non-overlapping cases. We further extend our technique to the 3-D case which also suffers from the same issues as 2-D detection. Extensive results on various public datasets (2-D/3-D, aerial/text/face images) with different base detectors show the effectiveness of our approach.
comment: 17 pages, 6 figures, 7 tables, accepted by ICLR 2023, TensorFlow code: https://github.com/yangxue0827/RotationDetection, PyTorch code: https://github.com/open-mmlab/mmrotate, Jittor code: https://github.com/Jittor/JDet
♻ ☆ Sample-efficient Surrogate Model for Frequency Response of Linear PDEs using Self-Attentive Complex Polynomials
Linear Partial Differential Equations (PDEs) govern the spatial-temporal dynamics of physical systems that are essential to building modern technology. When working with linear PDEs, designing a physical system for a specific outcome is difficult and costly due to slow and expensive explicit simulation of PDEs and the highly nonlinear relationship between a system's configuration and its behavior. In this work, we prove a parametric form that certain physical quantities in the Fourier domain must obey in linear PDEs, named the CZP (Constant-Zeros-Poles) framework. Applying CZP to antenna design, an industrial application using linear PDEs (i.e., Maxwell's equations), we derive a sample-efficient parametric surrogate model that directly predicts its scattering coefficients without explicit numerical PDE simulation. Combined with a novel image-based antenna representation and an attention-based neural network architecture, CZP outperforms baselines by 10% to 25% in terms of test loss and also is able to find 2D antenna designs verifiable by commercial software with $33\%$ greater success than baselines, when coupled with sequential search techniques like reinforcement learning.
♻ ☆ Surprising Instabilities in Training Deep Networks and a Theoretical Analysis
We discover restrained numerical instabilities in current training practices of deep networks with stochastic gradient descent (SGD). We show numerical error (on the order of the smallest floating point bit) induced from floating point arithmetic in training deep nets can be amplified significantly and result in significant test accuracy variance, comparable to the test accuracy variance due to stochasticity in SGD. We show how this is likely traced to instabilities of the optimization dynamics that are restrained, i.e., localized over iterations and regions of the weight tensor space. We do this by presenting a theoretical framework using numerical analysis of partial differential equations (PDE), and analyzing the gradient descent PDE of convolutional neural networks (CNNs). We show that it is stable only under certain conditions on the learning rate and weight decay. We show that rather than blowing up when the conditions are violated, the instability can be restrained. We show this is a consequence of the non-linear PDE associated with the gradient descent of the CNN, whose local linearization changes when over-driving the step size of the discretization, resulting in a stabilizing effect. We link restrained instabilities to the recently discovered Edge of Stability (EoS) phenomena, in which the stable step size predicted by classical theory is exceeded while continuing to optimize the loss and still converging. Because restrained instabilities occur at the EoS, our theory provides new predictions about the EoS, in particular, the role of regularization and the dependence on the network complexity.
♻ ☆ Revisiting Simple Regret: Fast Rates for Returning a Good Arm
Simple regret is a natural and parameter-free performance criterion for pure exploration in multi-armed bandits yet is less popular than the probability of missing the best arm or an $\epsilon$-good arm, perhaps due to lack of easy ways to characterize it. In this paper, we make significant progress on minimizing simple regret in both data-rich ($T\ge n$) and data-poor regime ($T \le n$) where $n$ is the number of arms, and $T$ is the number of samples. At its heart is our improved instance-dependent analysis of the well-known Sequential Halving (SH) algorithm, where we bound the probability of returning an arm whose mean reward is not within $\epsilon$ from the best (i.e., not $\epsilon$-good) for \textit{any} choice of $\epsilon>0$, although $\epsilon$ is not an input to SH. Our bound not only leads to an optimal worst-case simple regret bound of $\sqrt{n/T}$ up to logarithmic factors but also essentially matches the instance-dependent lower bound for returning an $\epsilon$-good arm reported by Katz-Samuels and Jamieson (2020). For the more challenging data-poor regime, we propose Bracketing SH (BSH) that enjoys the same improvement even without sampling each arm at least once. Our empirical study shows that BSH outperforms existing methods on real-world tasks.
♻ ☆ Efficient Privacy-Preserving Stochastic Nonconvex Optimization
While many solutions for privacy-preserving convex empirical risk minimization (ERM) have been developed, privacy-preserving nonconvex ERM remains a challenge. We study nonconvex ERM, which takes the form of minimizing a finite-sum of nonconvex loss functions over a training set. We propose a new differentially private stochastic gradient descent algorithm for nonconvex ERM that achieves strong privacy guarantees efficiently, and provide a tight analysis of its privacy and utility guarantees, as well as its gradient complexity. Our algorithm reduces gradient complexity while improves the best previous utility guarantee given by Wang et al. (NeurIPS 2017). Our experiments on benchmark nonconvex ERM problems demonstrate superior performance in terms of both training cost and utility gains compared with previous differentially private methods using the same privacy budgets.
comment: 29 pages, 5 figures, 3 tables. This version corrects a miscalculation in the previous proof, resulting in an improved utility bound for the algorithm
♻ ☆ SketchySGD: Reliable Stochastic Optimization via Robust Curvature Estimates
We introduce SketchySGD, a stochastic quasi-Newton method that uses sketching to approximate the curvature of the loss function. SketchySGD improves upon existing stochastic gradient methods in machine learning by using randomized low-rank approximations to the subsampled Hessian and by introducing an automated stepsize that works well across a wide range of convex machine learning problems. We show theoretically that SketchySGD with a fixed stepsize converges linearly to a small ball around the optimum. Further, in the ill-conditioned setting we show SketchySGD converges at a faster rate than SGD for least-squares problems. We validate this improvement empirically with ridge regression experiments on real data. Numerical experiments on both ridge and logistic regression problems show that SketchySGD can achieve comparable or better results to popular stochastic gradient methods with minimal hyperparameter tuning. The robustness of SketchySGD to hyperparameters is an advantage over other stochastic gradient methods, most of which require careful hyperparameter tuning (especially of the learning rate) to obtain good performance.
comment: 20 pages, 10 figures, 4 tables
♻ ☆ Self-Programming Artificial Intelligence Using Code-Generating Language Models
Recent progress in large-scale language models has enabled breakthroughs in previously intractable computer programming tasks. Prior work in meta-learning and neural architecture search has led to substantial successes across various task domains, spawning myriad approaches for algorithmically optimizing the design and learning dynamics of deep learning models. At the intersection of these research areas, we implement a code-generating language model with the ability to modify its own source code. Self-programming AI algorithms have been of interest since the dawn of AI itself. Although various theoretical formulations of generalized self-programming AI have been posed, no such system has been successfully implemented to date under real-world computational constraints. Applying AI-based code generation to AI itself, we develop and experimentally validate the first practical implementation of a self-programming AI system. We empirically show that a self-programming AI implemented using a code generation model can successfully modify its own source code to improve performance and program sub-models to perform auxiliary tasks. Our model can self-modify various properties including model architecture, computational capacity, and learning dynamics.
♻ ☆ Beyond Invariance: Test-Time Label-Shift Adaptation for Distributions with "Spurious" Correlations
Spurious correlations, or correlations that change across domains where a model can be deployed, present significant challenges to real-world applications of machine learning models. However, such correlations are not always "spurious"; often, they provide valuable prior information for a prediction. Here, we present a test-time adaptation method that exploits the spurious correlation phenomenon, in contrast to recent approaches that attempt to eliminate spurious correlations through invariance. We consider situations where the prior distribution $p(y, z)$, which models the dependence between the class label $y$ and the "nuisance" factors $z$, may change across domains, but the generative model for features $p(\mathbf{x}|y, z)$ is constant. We note that this corresponds to an expanded version of the label shift assumption, where the labels now also include the nuisance factors $z$. Based on this observation, we train a classifier to predict $p(y, z|\mathbf{x})$ on the source distribution, and propose a test-time label shift correction that adapts to changes in the marginal distribution $p(y, z)$ using unlabeled samples from the target domain. We evaluate our method, which we call "Test-Time Label-Shift Adaptation" (TTLSA), on two different image datasets -- the CheXpert chest X-ray dataset and the Colored MNIST dataset -- and show a significant improvement over baseline methods. Code reproducing experiments is available at https://github.com/nalzok/test-time-label-shift .
comment: 17 pages, 7 figures
♻ ☆ Statistical treatment of convolutional neural network super-resolution of inland surface wind for subgrid-scale variability quantification
Machine learning models have been employed to perform either physics-free data-driven or hybrid dynamical downscaling of climate data. Most of these implementations operate over relatively small downscaling factors because of the challenge of recovering fine-scale information from coarse data. This limits their compatibility with many global climate model outputs, often available between $\sim$50--100 km resolution, to scales of interest such as cloud resolving or urban scales. This study systematically examines the capability of convolutional neural networks (CNNs) to downscale surface wind speed data over land surface from different coarse resolutions (25 km, 48 km, and 100 km resolution) to 3 km. For each downscaling factor, we consider three CNN configurations that generate super-resolved predictions of fine-scale wind speed, which take between 1 to 3 input fields: coarse wind speed, fine-scale topography, and diurnal cycle. In addition to fine-scale wind speeds, probability density function parameters are generated, through which sample wind speeds can be generated accounting for the intrinsic stochasticity of wind speed. For generalizability assessment, CNN models are tested on regions with different topography and climate that are unseen during training. The evaluation of super-resolved predictions focuses on subgrid-scale variability and the recovery of extremes. Models with coarse wind and fine topography as inputs exhibit the best performance compared with other model configurations, operating across the same downscaling factor. Our diurnal cycle encoding results in lower out-of-sample generalizability compared with other input configurations.
♻ ☆ LEAF: Navigating Concept Drift in Cellular Networks
Operational networks commonly rely on machine learning models for many tasks, including detecting anomalies, inferring application performance, and forecasting demand. Yet, model accuracy can degrade due to concept drift, whereby the relationship between the features and the target to be predicted changes. Mitigating concept drift is an essential part of operationalizing machine learning models in general, but is of particular importance in networking's highly dynamic deployment environments. In this paper, we first characterize concept drift in a large cellular network for a major metropolitan area in the United States. We find that concept drift occurs across many important key performance indicators (KPIs), independently of the model, training set size, and time interval -- thus necessitating practical approaches to detect, explain, and mitigate it. We then show that frequent model retraining with newly available data is not sufficient to mitigate concept drift, and can even degrade model accuracy further. Finally, we develop a new methodology for concept drift mitigation, Local Error Approximation of Features (LEAF). LEAF works by detecting drift; explaining the features and time intervals that contribute the most to drift; and mitigates it using forgetting and over-sampling. We evaluate LEAF against industry-standard mitigation approaches (notably, periodic retraining) with more than four years of cellular KPI data. Our initial tests with a major cellular provider in the US show that LEAF consistently outperforms periodic and triggered retraining on complex, real-world data while reducing costly retraining operations.
♻ ☆ Distributionally Robust Causal Inference with Observational Data
We consider the estimation of average treatment effects in observational studies and propose a new framework of robust causal inference with unobserved confounders. Our approach is based on distributionally robust optimization and proceeds in two steps. We first specify the maximal degree to which the distribution of unobserved potential outcomes may deviate from that of observed outcomes. We then derive sharp bounds on the average treatment effects under this assumption. Our framework encompasses the popular marginal sensitivity model as a special case, and we demonstrate how the proposed methodology can address a primary challenge of the marginal sensitivity model that it produces uninformative results when unobserved confounders substantially affect treatment and outcome. Specifically, we develop an alternative sensitivity model, called the distributional sensitivity model, under the assumption that heterogeneity of treatment effect due to unobserved variables is relatively small. Unlike the marginal sensitivity model, the distributional sensitivity model allows for potential lack of overlap and often produces informative bounds even when unobserved variables substantially affect both treatment and outcome. Finally, we show how to extend the distributional sensitivity model to difference-in-differences designs and settings with instrumental variables. Through simulation and empirical studies, we demonstrate the applicability of the proposed methodology.
♻ ☆ Towards Optimal Branching of Linear and Semidefinite Relaxations for Neural Network Robustness Certification
In this paper, we study certifying the robustness of ReLU neural networks against adversarial input perturbations. To diminish the relaxation error suffered by the popular linear programming (LP) and semidefinite programming (SDP) certification methods, we take a branch-and-bound approach to propose partitioning the input uncertainty set and solving the relaxations on each part separately. We show that this approach reduces relaxation error, and that the error is eliminated entirely upon performing an LP relaxation with a partition intelligently designed to exploit the nature of the ReLU activations. To scale this approach to large networks, we consider using a coarser partition whereby the number of parts in the partition is reduced. We prove that computing such a coarse partition that directly minimizes the LP relaxation error is NP-hard. By instead minimizing the worst-case LP relaxation error, we develop a closed-form branching scheme. We extend the analysis to the SDP, where the feasible set geometry is exploited to design a branching scheme that minimizes the worst-case SDP relaxation error. Experiments on MNIST, CIFAR-10, and Wisconsin breast cancer diagnosis classifiers demonstrate significant increases in the percentages of test samples certified. By independently increasing the input size and the number of layers, we empirically illustrate under which regimes the branched LP and branched SDP are best applied.
comment: This is an extension of our IEEE CDC 2020 conference paper arXiv:2004.00570
Multimedia 4
☆ Introduction to Arithmetic Coding -- Theory and Practice
This introduction to arithmetic coding is divided in two parts. The first explains how and why arithmetic coding works. We start presenting it in very general terms, so that its simplicity is not lost under layers of implementation details. Next, we show some of its basic properties, which are later used in the computational techniques required for a practical implementation. In the second part, we cover the practical implementation aspects, including arithmetic operations with low precision, the subdivision of coding and modeling, and the realization of adaptive encoders. We also analyze the arithmetic coding computational complexity, and techniques to reduce it.
comment: Hewlett-Packard Laboratories Report; Chapter in Lossless Compression Handbook (ed. K. Sayood), Academic Press
☆ Scene2BIR: Material-aware learning-based binaural impulse response generator for reconstructed real-world 3D scenes
We present an end-to-end binaural impulse response generator (BIR) to generate plausible sounds in real-time for real-world models. Our approach uses a novel neural-network-based BIR generator (Scene2BIR) for the reconstructed 3D model. We propose a graph neural network that uses both the material and the topology information of the 3D scenes and generates a scene latent vector. Moreover, we use a conditional generative adversarial network (CGAN) to generate BIRs from the scene latent vector. Our network is able to handle holes or other artifacts in the reconstructed 3D mesh model. We present an efficient cost function to the generator network to incorporate spatial audio effects. Given the source and the listener position, our approach can generate a BIR in 0.1 milliseconds on an NVIDIA GeForce RTX 2080 Ti GPU and can easily handle multiple sources. We have evaluated the accuracy of our approach with real-world captured BIRs and an interactive geometric sound propagation algorithm.
comment: Project page: https://anton-jeran.github.io/S2BIR/
☆ MMRec: Simplifying Multimodal Recommendation
This paper presents an open-source toolbox, MMRec for multimodal recommendation. MMRec simplifies and canonicalizes the process of implementing and comparing multimodal recommendation models. The objective of MMRec is to provide a unified and configurable arena that can minimize the effort in implementing and testing multimodal recommendation models. It enables multimodal models, ranging from traditional matrix factorization to modern graph-based algorithms, capable of fusing information from multiple modalities simultaneously. Our documentation, examples, and source code are available at \url{https://github.com/enoche/MMRec}.
comment: 3 pages
♻ ☆ Leveraging Modality-specific Representations for Audio-visual Speech Recognition via Reinforcement Learning AAAI2023
Audio-visual speech recognition (AVSR) has gained remarkable success for ameliorating the noise-robustness of speech recognition. Mainstream methods focus on fusing audio and visual inputs to obtain modality-invariant representations. However, such representations are prone to over-reliance on audio modality as it is much easier to recognize than video modality in clean conditions. As a result, the AVSR model underestimates the importance of visual stream in face of noise corruption. To this end, we leverage visual modality-specific representations to provide stable complementary information for the AVSR task. Specifically, we propose a reinforcement learning (RL) based framework called MSRL, where the agent dynamically harmonizes modality-invariant and modality-specific representations in the auto-regressive decoding process. We customize a reward function directly related to task-specific metrics (i.e., word error rate), which encourages the MSRL to effectively explore the optimal integration strategy. Experimental results on the LRS3 dataset show that the proposed method achieves state-of-the-art in both clean and various noisy conditions. Furthermore, we demonstrate the better generality of MSRL system than other baselines when test set contains unseen noises.
comment: Accepted by AAAI2023
Computation and Language 40
☆ Improving Few-Shot Generalization by Exploring and Exploiting Auxiliary Data
Few-shot learning involves learning an effective model from only a few labeled datapoints. The use of a small training set makes it difficult to avoid overfitting but also makes few-shot learning applicable to many important real-world settings. In this work, we focus on Few-shot Learning with Auxiliary Data (FLAD), a training paradigm that assumes access to auxiliary data during few-shot learning in hopes of improving generalization. Introducing auxiliary data during few-shot learning leads to essential design choices where hand-designed heuristics can lead to sub-optimal performance. In this work, we focus on automated sampling strategies for FLAD and relate them to the explore-exploit dilemma that is central in multi-armed bandit settings. Based on this connection we propose two algorithms -- EXP3-FLAD and UCB1-FLAD -- and compare them with methods that either explore or exploit, finding that the combination of exploration and exploitation is crucial. Using our proposed algorithms to train T5 yields a 9% absolute improvement over the explicitly multi-task pre-trained T0 model across 11 datasets.
comment: 19 pages, 7 figures, code available at https://github.com/alon-albalak/FLAD
☆ 'Generative CI' through Collective Response Systems
How can many people (who may disagree) come together to answer a question or make a decision? "Collective response systems" are a type of generative collective intelligence (CI) facilitation process meant to address this challenge. They enable a form of "generative voting", where both the votes, and the choices of what to vote on, are provided by the group. Such systems overcome the traditional limitations of polling, town halls, standard voting, referendums, etc. The generative CI outputs of collective response systems can also be chained together into iterative "collective dialogues", analogously to some kinds of generative AI. Technical advances across domains including recommender systems, language models, and human-computer interaction have led to the development of innovative and scalable collective response systems. For example, Polis has been used around the world to support policy-making at different levels of government, and Remesh has been used by the UN to understand the challenges and needs of ordinary people across war-torn countries. This paper aims to develop a shared language by defining the structure, processes, properties, and principles of such systems. Collective response systems allow non-confrontational exploration of divisive issues, help identify common ground, and elicit insights from those closest to the issues. As a result, they can help overcome gridlock around conflict and governance challenges, increase trust, and develop mandates. Continued progress toward their development and adoption could help revitalize democracies, reimagine corporate governance, transform conflict, and govern powerful AI systems -- both as a complement to deeper deliberative democratic processes and as an option where deeper processes are not applicable or possible.
comment: 5 pages
☆ Does Vision Accelerate Hierarchical Generalization of Neural Language Learners?
Neural language models (LMs) are arguably less data-efficient than humans -- why does this gap occur? In this study, we hypothesize that this gap stems from the learners' accessibility to modalities other than text, specifically, vision. We conducted two complementary experiments (using noisy, realistic data and a simplified, artificial one) toward the advantage of vision in the syntactic generalization of LMs. Our results showed that vision accelerated a proper linguistic generalization in the simplified, artificial setting, but LMs struggled with the noisy, realistic setting. These mixed results indicate several possibilities, e.g., vision can potentially boost language acquisition, but learners' additional visual/linguistic prior knowledge should be needed to robustly make use of raw images for efficient language acquisition.
comment: Work in progress
☆ Are UD Treebanks Getting More Consistent? A Report Card for English UD
Recent efforts to consolidate guidelines and treebanks in the Universal Dependencies project raise the expectation that joint training and dataset comparison is increasingly possible for high-resource languages such as English, which have multiple corpora. Focusing on the two largest UD English treebanks, we examine progress in data consolidation and answer several questions: Are UD English treebanks becoming more internally consistent? Are they becoming more like each other and to what extent? Is joint training a good idea, and if so, since which UD version? Our results indicate that while consolidation has made progress, joint models may still suffer from inconsistencies, which hamper their ability to leverage a larger pool of training data.
comment: Proceedings of the Sixth Workshop on Universal Dependencies (UDW 2023)
☆ Synthetic Prompting: Generating Chain-of-Thought Demonstrations for Large Language Models
Large language models can perform various reasoning tasks by using chain-of-thought prompting, which guides them to find answers through step-by-step demonstrations. However, the quality of the prompts depends on the demonstrations given to the models, and creating many of them by hand is costly. We introduce Synthetic prompting, a method that leverages a few handcrafted examples to prompt the model to generate more examples by itself, and selects effective demonstrations to elicit better reasoning. Our method alternates between a backward and forward process to generate new examples. The backward process generates a question that match a sampled reasoning chain, so that the question is solvable and clear. The forward process produces a more detailed reasoning chain for the question, improving the quality of the example. We evaluate our method on numerical, symbolic, and algorithmic reasoning tasks, and show that it outperforms existing prompting techniques.
comment: Preprint
☆ Zero Shot Transfer of Legal Judgement Prediction as Article-aware Entailment for the European Court of Human Rights EACL
In this paper, we cast Legal Judgment Prediction (LJP) from text on European Court of Human Rights cases as an entailment task, where the case outcome is classified from a combined input of case facts and convention articles. This configuration facilitates the model learning legal reasoning ability in mapping article text to specific fact text. It also provides the opportunity to evaluate the model's ability to generalize to zero-shot settings when asked to classify the case outcome with respect to articles not seen during training. We devise zero-shot LJP experiments and apply domain adaptation methods based on domain discriminator and Wasserstein distance. Our results demonstrate that the entailment architecture outperforms straightforward fact classification. We also find that domain adaptation methods improve zero-shot transfer performance, with article relatedness and encoder pre-training influencing the effect.
comment: EACL Findings 2023
☆ The RW3D: A multi-modal panel dataset to understand the psychological impact of the pandemic
Besides far-reaching public health consequences, the COVID-19 pandemic had a significant psychological impact on people around the world. To gain further insight into this matter, we introduce the Real World Worry Waves Dataset (RW3D). The dataset combines rich open-ended free-text responses with survey data on emotions, significant life events, and psychological stressors in a repeated-measures design in the UK over three years (2020: n=2441, 2021: n=1716 and 2022: n=1152). This paper provides background information on the data collection procedure, the recorded variables, participants' demographics, and higher-order psychological and text-based derived variables that emerged from the data. The RW3D is a unique primary data resource that could inspire new research questions on the psychological impact of the pandemic, especially those that connect modalities (here: text data, psychological survey variables and demographics) over time.
comment: preprint
☆ Co-Writing with Opinionated Language Models Affects Users' Views
If large language models like GPT-3 preferably produce a particular point of view, they may influence people's opinions on an unknown scale. This study investigates whether a language-model-powered writing assistant that generates some opinions more often than others impacts what users write - and what they think. In an online experiment, we asked participants (N=1,506) to write a post discussing whether social media is good for society. Treatment group participants used a language-model-powered writing assistant configured to argue that social media is good or bad for society. Participants then completed a social media attitude survey, and independent judges (N=500) evaluated the opinions expressed in their writing. Using the opinionated language model affected the opinions expressed in participants' writing and shifted their opinions in the subsequent attitude survey. We discuss the wider implications of our results and argue that the opinions built into AI language technologies need to be monitored and engineered more carefully.
☆ Exploring Semantic Perturbations on Grover
With news and information being as easy to access as they currently are, it is more important than ever to ensure that people are not mislead by what they read. Recently, the rise of neural fake news (AI-generated fake news) and its demonstrated effectiveness at fooling humans has prompted the development of models to detect it. One such model is the Grover model, which can both detect neural fake news to prevent it, and generate it to demonstrate how a model could be misused to fool human readers. In this work we explore the Grover model's fake news detection capabilities by performing targeted attacks through perturbations on input news articles. Through this we test Grover's resilience to these adversarial attacks and expose some potential vulnerabilities which should be addressed in further iterations to ensure it can detect all types of fake news accurately.
comment: 15 pages, 12 figures, 1 table, capstone research in machine learning
☆ You Are What You Talk About: Inducing Evaluative Topics for Personality Analysis EMNLP 2022
Expressing attitude or stance toward entities and concepts is an integral part of human behavior and personality. Recently, evaluative language data has become more accessible with social media's rapid growth, enabling large-scale opinion analysis. However, surprisingly little research examines the relationship between personality and evaluative language. To bridge this gap, we introduce the notion of evaluative topics, obtained by applying topic models to pre-filtered evaluative text from social media. We then link evaluative topics to individual text authors to build their evaluative profiles. We apply evaluative profiling to Reddit comments labeled with personality scores and conduct an exploratory study on the relationship between evaluative topics and Big Five personality facets, aiming for a more interpretable, facet-level analysis. Finally, we validate our approach by observing correlations consistent with prior research in personality psychology.
comment: Accepted at EMNLP 2022 (Findings), NLP+CSS
☆ Feed-Forward Blocks Control Contextualization in Masked Language Models
Understanding the inner workings of neural network models is a crucial step for rationalizing their output and refining their architecture. Transformer-based models are the core of recent natural language processing and have been analyzed typically with attention patterns as their epoch-making feature is contextualizing surrounding input words via attention mechanisms. In this study, we analyze their inner contextualization by considering all the components, including the feed-forward block (i.e., a feed-forward layer and its surrounding residual and normalization layers) as well as the attention. Our experiments with masked language models show that each of the previously overlooked components did modify the degree of the contextualization in case of processing special word-word pairs (e.g., consisting of named entities). Furthermore, we find that some components cancel each other's effects. Our results could update the typical view about each component's roles (e.g., attention performs contextualization, and the other components serve different roles) in the Transformer layer.
comment: 13 pages, 15 figures
☆ HunSum-1: an Abstractive Summarization Dataset for Hungarian
We introduce HunSum-1: a dataset for Hungarian abstractive summarization, consisting of 1.14M news articles. The dataset is built by collecting, cleaning and deduplicating data from 9 major Hungarian news sites through CommonCrawl. Using this dataset, we build abstractive summarizer models based on huBERT and mT5. We demonstrate the value of the created dataset by performing a quantitative and qualitative analysis on the models' results. The HunSum-1 dataset, all models used in our experiments and our code are available open source.
☆ Improved Knowledge Distillation for Pre-trained Language Models via Knowledge Selection EMNLP
Knowledge distillation addresses the problem of transferring knowledge from a teacher model to a student model. In this process, we typically have multiple types of knowledge extracted from the teacher model. The problem is to make full use of them to train the student model. Our preliminary study shows that: (1) not all of the knowledge is necessary for learning a good student model, and (2) knowledge distillation can benefit from certain knowledge at different training steps. In response to these, we propose an actor-critic approach to selecting appropriate knowledge to transfer during the process of knowledge distillation. In addition, we offer a refinement of the training algorithm to ease the computational burden. Experimental results on the GLUE datasets show that our method outperforms several strong knowledge distillation baselines significantly.
comment: accepted by EMNLP (Findings) 2022
☆ KNNs of Semantic Encodings for Rating Prediction
This paper explores a novel application of textual semantic similarity to user-preference representation for rating prediction. The approach represents a user's preferences as a graph of textual snippets from review text, where the edges are defined by semantic similarity. This textual, memory-based approach to rating prediction enables review-based explanations for recommendations. The method is evaluated quantitatively, highlighting that leveraging text in this way outperforms both strong memory-based and model-based collaborative filtering baselines.
☆ On the Role of Morphological Information for Contextual Lemmatization
Lemmatization is a Natural Language Processing (NLP) task which consists of producing, from a given inflected word, its canonical form or lemma. Lemmatization is one of the basic tasks that facilitate downstream NLP applications, and is of particular importance for high-inflected languages. Given that the process to obtain a lemma from an inflected word can be explained by looking at its morphosyntactic category, including fine-grained morphosyntactic information to train contextual lemmatizers has become common practice, without analyzing whether that is the optimum in terms of downstream performance. Thus, in this paper we empirically investigate the role of morphological information to develop contextual lemmatizers in six languages within a varied spectrum of morphological complexity: Basque, Turkish, Russian, Czech, Spanish and English. Furthermore, and unlike the vast majority of previous work, we also evaluate lemmatizers in out-of-domain settings, which constitutes, after all, their most common application use. The results of our study are rather surprising: (i) providing lemmatizers with fine-grained morphological features during training is not that beneficial, not even for agglutinative languages; (ii) in fact, modern contextual word representations seem to implicitly encode enough morphological information to obtain good contextual lemmatizers without seeing any explicit morphological signal; (iii) the best lemmatizers out-of-domain are those using simple UPOS tags or those trained without morphology; (iv) current evaluation practices for lemmatization are not adequate to clearly discriminate between models.
comment: 24 pages, 6 figures, 6 tables
☆ mPLUG-2: A Modularized Multi-modal Foundation Model Across Text, Image and Video
Recent years have witnessed a big convergence of language, vision, and multi-modal pretraining. In this work, we present mPLUG-2, a new unified paradigm with modularized design for multi-modal pretraining, which can benefit from modality collaboration while addressing the problem of modality entanglement. In contrast to predominant paradigms of solely relying on sequence-to-sequence generation or encoder-based instance discrimination, mPLUG-2 introduces a multi-module composition network by sharing common universal modules for modality collaboration and disentangling different modality modules to deal with modality entanglement. It is flexible to select different modules for different understanding and generation tasks across all modalities including text, image, and video. Empirical study shows that mPLUG-2 achieves state-of-the-art or competitive results on a broad range of over 30 downstream tasks, spanning multi-modal tasks of image-text and video-text understanding and generation, and uni-modal tasks of text-only, image-only, and video-only understanding. Notably, mPLUG-2 shows new state-of-the-art results of 48.0 top-1 accuracy and 80.3 CIDEr on the challenging MSRVTT video QA and video caption tasks with a far smaller model size and data scale. It also demonstrates strong zero-shot transferability on vision-language and video-language tasks. Code and models will be released in https://github.com/alibaba/AliceMind.
☆ An Empirical Study on the Transferability of Transformer Modules in Parameter-Efficient Fine-Tuning EMNLP 2022
Parameter-efficient fine-tuning approaches have recently garnered a lot of attention. Having considerably lower number of trainable weights, these methods can bring about scalability and computational effectiveness. In this paper, we look for optimal sub-networks and investigate the capability of different transformer modules in transferring knowledge from a pre-trained model to a downstream task. Our empirical results suggest that every transformer module in BERT can act as a winning ticket: fine-tuning each specific module while keeping the rest of the network frozen can lead to comparable performance to the full fine-tuning. Among different modules, LayerNorms exhibit the best capacity for knowledge transfer with limited trainable weights, to the extent that, with only 0.003% of all parameters in the layer-wise analysis, they show acceptable performance on various target tasks. On the reasons behind their effectiveness, we argue that their notable performance could be attributed to their high-magnitude weights compared to that of the other modules in the pre-trained BERT.
comment: Accepted at EMNLP 2022 (main conference)
☆ Attention Link: An Efficient Attention-Based Low Resource Machine Translation Architecture
Transformers have achieved great success in machine translation, but transformer-based NMT models often require millions of bilingual parallel corpus for training. In this paper, we propose a novel architecture named as attention link (AL) to help improve transformer models' performance, especially in low training resources. We theoretically demonstrate the superiority of our attention link architecture in low training resources. Besides, we have done a large number of experiments, including en-de, de-en, en-fr, en-it, it-en, en-ro translation tasks on the IWSLT14 dataset as well as real low resources scene on bn-gu and gu-ta translation tasks on the CVIT PIB dataset. All the experiment results show our attention link is powerful and can lead to a significant improvement. In addition, we achieve a 37.9 BLEU score, a new sota, on the IWSLT14 de-en task by combining our attention link and other advanced methods.
☆ Evaluating TCFD Reporting: A New Application of Zero-Shot Analysis to Climate-Related Financial Disclosures
We examine climate-related disclosures in a large sample of reports published by banks that officially endorsed the recommendations of the Task Force for Climate-related Financial Disclosures (TCFD). In doing so, we introduce a new application of the zero-shot text classification. By developing a set of fine-grained TCFD labels, we show that zero-shot analysis is a useful tool for classifying climate-related disclosures without further model training. Overall, our findings indicate that corporate climate-related disclosures grew dynamically after the launch of the TCFD recommendations. However, there are marked differences in the extent of reporting by recommended disclosure topic, suggesting that some recommendations have not yet been fully met. Our findings yield important conclusions for the design of climate-related disclosure frameworks.
☆ An Evaluation of Persian-English Machine Translation Datasets with Transformers
Nowadays, many researchers are focusing their attention on the subject of machine translation (MT). However, Persian machine translation has remained unexplored despite a vast amount of research being conducted in languages with high resources, such as English. Moreover, while a substantial amount of research has been undertaken in statistical machine translation for some datasets in Persian, there is currently no standard baseline for transformer-based text2text models on each corpus. This study collected and analysed the most popular and valuable parallel corpora, which were used for Persian-English translation. Furthermore, we fine-tuned and evaluated two state-of-the-art attention-based seq2seq models on each dataset separately (48 results). We hope this paper will assist researchers in comparing their Persian to English and vice versa machine translation results to a standard baseline.
☆ Filtering Context Mitigates Scarcity and Selection Bias in Political Ideology Prediction
We propose a novel supervised learning approach for political ideology prediction (PIP) that is capable of predicting out-of-distribution inputs. This problem is motivated by the fact that manual data-labeling is expensive, while self-reported labels are often scarce and exhibit significant selection bias. We propose a novel statistical model that decomposes the document embeddings into a linear superposition of two vectors; a latent neutral \emph{context} vector independent of ideology, and a latent \emph{position} vector aligned with ideology. We train an end-to-end model that has intermediate contextual and positional vectors as outputs. At deployment time, our model predicts labels for input documents by exclusively leveraging the predicted positional vectors. On two benchmark datasets we show that our model is capable of outputting predictions even when trained with as little as 5\% biased data, and is significantly more accurate than the state-of-the-art. Through crowd-sourcing we validate the neutrality of contextual vectors, and show that context filtering results in ideological concentration, allowing for prediction on out-of-distribution examples.
☆ A Transaction Represented with Weighted Finite-State Transducers
Not all contracts are good, but all good contracts can be expressed as a finite-state transition system ("State-Transition Contracts"). Contracts that can be represented as State-Transition Contracts discretize fat-tailed risk to foreseeable, managed risk, define the boundary of relevant events governed by the relationship, and eliminate the potential of inconsistent contractual provisions. Additionally, State-Transition Contracts reap the substantial benefit of being able to be analyzed under the rules governing the science of the theory of computation. Simple State-Transition Contracts can be represented as discrete finite automata; more complicated State-Transition Contracts, such as those that have downstream effects on other agreements or complicated pathways of performance, benefit from representation as weighted finite-state transducers, with weights assigned as costs, penalties, or probabilities of transitions. This research paper (the "Research" or "Paper") presents a complex legal transaction represented as weighted finite-state transducers. Furthermore, we show that the mathematics/algorithms permitted by the algebraic structure of weighted finite-state transducers provides actionable, legal insight into the transaction.
comment: 2 figures, 3 tables, 2 appendices, Recognition Technologies, Inc. Technical Report
☆ Detecting Lexical Borrowings from Dominant Languages in Multilingual Wordlists
Language contact is a pervasive phenomenon reflected in the borrowing of words from donor to recipient languages. Most computational approaches to borrowing detection treat all languages under study as equally important, even though dominant languages have a stronger impact on heritage languages than vice versa. We test new methods for lexical borrowing detection in contact situations where dominant languages play an important role, applying two classical sequence comparison methods and one machine learning method to a sample of seven Latin American languages which have all borrowed extensively from Spanish. All methods perform well, with the supervised machine learning system outperforming the classical systems. A review of detection errors shows that borrowing detection could be substantially improved by taking into account donor words with divergent meanings from recipient words.
comment: To appear at The 17th Conference of the European Chapter of the Association for Computational Linguistics. See https://www.aclweb.org/portal/content/17th-conference-european-chapter-association-computational-linguistics
☆ Program Generation from Diverse Video Demonstrations
The ability to use inductive reasoning to extract general rules from multiple observations is a vital indicator of intelligence. As humans, we use this ability to not only interpret the world around us, but also to predict the outcomes of the various interactions we experience. Generalising over multiple observations is a task that has historically presented difficulties for machines to grasp, especially when requiring computer vision. In this paper, we propose a model that can extract general rules from video demonstrations by simultaneously performing summarisation and translation. Our approach differs from prior works by framing the problem as a multi-sequence-to-sequence task, wherein summarisation is learnt by the model. This allows our model to utilise edge cases that would otherwise be suppressed or discarded by traditional summarisation techniques. Additionally, we show that our approach can handle noisy specifications without the need for additional filtering methods. We evaluate our model by synthesising programs from video demonstrations in the Vizdoom environment achieving state-of-the-art results with a relative increase of 11.75% program accuracy on prior works
☆ Unsupervised Entity Alignment for Temporal Knowledge Graphs WWW
Entity alignment (EA) is a fundamental data integration task that identifies equivalent entities between different knowledge graphs (KGs). Temporal Knowledge graphs (TKGs) extend traditional knowledge graphs by introducing timestamps, which have received increasing attention. State-of-the-art time-aware EA studies have suggested that the temporal information of TKGs facilitates the performance of EA. However, existing studies have not thoroughly exploited the advantages of temporal information in TKGs. Also, they perform EA by pre-aligning entity pairs, which can be labor-intensive and thus inefficient. In this paper, we present DualMatch which effectively fuses the relational and temporal information for EA. DualMatch transfers EA on TKGs into a weighted graph matching problem. More specifically, DualMatch is equipped with an unsupervised method, which achieves EA without necessitating seed alignment. DualMatch has two steps: (i) encoding temporal and relational information into embeddings separately using a novel label-free encoder, Dual-Encoder; and (ii) fusing both information and transforming it into alignment using a novel graph-matching-based decoder, GM-Decoder. DualMatch is able to perform EA on TKGs with or without supervision, due to its capability of effectively capturing temporal information. Extensive experiments on three real-world TKG datasets offer the insight that DualMatch outperforms the state-of-the-art methods in terms of H@1 by 2.4% - 10.7% and MRR by 1.7% - 7.6%, respectively.
comment: Accepted by The Web Conference (WWW) 2023 Research Track
☆ User Study for Improving Tools for Bible Translation
Technology has increasingly become an integral part of the Bible translation process. Over time, both the translation process and relevant technology have evolved greatly. More recently, the field of Natural Language Processing (NLP) has made great progress in solving some problems previously thought impenetrable. Through this study we endeavor to better understand and communicate about a segment of the current landscape of the Bible translation process as it relates to technology and identify pertinent issues. We conduct several interviews with individuals working in different levels of the Bible translation process from multiple organizations to identify gaps and bottlenecks where technology (including recent advances in AI) could potentially play a pivotal role in reducing translation time and improving overall quality.
comment: 5 pages, Report of user-study
☆ Leveraging task dependency and contrastive learning for Legal Judgement Prediction on the European Court of Human Rights EACL 2023
We report on an experiment in legal judgement prediction on European Court of Human Rights cases where our model first learns to predict the convention articles allegedly violated by the state from case facts descriptions, and subsequently utilizes that information to predict a finding of a violation by the court. We assess the dependency between these two tasks at the feature and outcome level. Furthermore, we leverage a hierarchical contrastive loss to pull together article specific representations of cases at the higher level level, leading to distinctive article clusters, and further pulls the cases in each article cluster based on their outcome leading to sub-clusters of cases with similar outcomes. Our experiment results demonstrate that, given a static pre-trained encoder, our models produce a small but consistent improvement in prediction performance over single-task and joint models without contrastive loss.
comment: EACL 2023
☆ Visually Grounded Keyword Detection and Localisation for Low-Resource Languages
This study investigates the use of Visually Grounded Speech (VGS) models for keyword localisation in speech. The study focusses on two main research questions: (1) Is keyword localisation possible with VGS models and (2) Can keyword localisation be done cross-lingually in a real low-resource setting? Four methods for localisation are proposed and evaluated on an English dataset, with the best-performing method achieving an accuracy of 57%. A new dataset containing spoken captions in Yoruba language is also collected and released for cross-lingual keyword localisation. The cross-lingual model obtains a precision of 16% in actual keyword localisation and this performance can be improved by initialising from a model pretrained on English data. The study presents a detailed analysis of the model's success and failure modes and highlights the challenges of using VGS models for keyword localisation in low-resource settings.
comment: PhD dissertation, University of Stellenbosch, 108 pages, submitted and accepted 2023
☆ Collaborating with language models for embodied reasoning NeurIPS 2022
Reasoning in a complex and ambiguous environment is a key goal for Reinforcement Learning (RL) agents. While some sophisticated RL agents can successfully solve difficult tasks, they require a large amount of training data and often struggle to generalize to new unseen environments and new tasks. On the other hand, Large Scale Language Models (LSLMs) have exhibited strong reasoning ability and the ability to to adapt to new tasks through in-context learning. However, LSLMs do not inherently have the ability to interrogate or intervene on the environment. In this work, we investigate how to combine these complementary abilities in a single system consisting of three parts: a Planner, an Actor, and a Reporter. The Planner is a pre-trained language model that can issue commands to a simple embodied agent (the Actor), while the Reporter communicates with the Planner to inform its next command. We present a set of tasks that require reasoning, test this system's ability to generalize zero-shot and investigate failure cases, and demonstrate how components of this system can be trained with reinforcement-learning to improve performance.
comment: Presented at NeurIPS 2022 Language and Reinforcement Learning Workshop (best paper) and NeurIPS 2022 Foundation Models for Decision Making Workshop. 4 pages main; 14 pages total (including references and appendix); 3 figures
☆ AmbiCoref: Evaluating Human and Model Sensitivity to Ambiguous Coreference EACL 2023
Given a sentence "Abby told Brittney that she upset Courtney", one would struggle to understand who "she" refers to, and ask for clarification. However, if the word "upset" were replaced with "hugged", "she" unambiguously refers to Abby. We study if modern coreference resolution models are sensitive to such pronominal ambiguity. To this end, we construct AmbiCoref, a diagnostic corpus of minimal sentence pairs with ambiguous and unambiguous referents. Our examples generalize psycholinguistic studies of human perception of ambiguity around particular arrangements of verbs and their arguments. Analysis shows that (1) humans are less sure of referents in ambiguous AmbiCoref examples than unambiguous ones, and (2) most coreference models show little difference in output between ambiguous and unambiguous pairs. We release AmbiCoref as a diagnostic corpus for testing whether models treat ambiguity similarly to human.
comment: EACL 2023 Findings
☆ Inference of Partial Colexifications from Multilingual Wordlists
The past years have seen a drastic rise in studies devoted to the investigation of colexification patterns in individual languages families in particular and the languages of the world in specific. Specifically computational studies have profited from the fact that colexification as a scientific construct is easy to operationalize, enabling scholars to infer colexification patterns for large collections of cross-linguistic data. Studies devoted to partial colexifications -- colexification patterns that do not involve entire words, but rather various parts of words--, however, have been rarely conducted so far. This is not surprising, since partial colexifications are less easy to deal with in computational approaches and may easily suffer from all kinds of noise resulting from false positive matches. In order to address this problem, this study proposes new approaches to the handling of partial colexifications by (1) proposing new models with which partial colexification patterns can be represented, (2) developing new efficient methods and workflows which help to infer various types of partial colexification patterns from multilingual wordlists, and (3) illustrating how inferred patterns of partial colexifications can be computationally analyzed and interactively visualized.
☆ DANES: Deep Neural Network Ensemble Architecture for Social and Textual Context-aware Fake News Detection
The growing popularity of social media platforms has simplified the creation and distribution of news articles but also creates a conduit for spreading fake news. In consequence, the need arises for effective context-aware fake news detection mechanisms, where the contextual information can be built either from the textual content of posts or from available social data (e.g., information about the users, reactions to posts, or the social network). In this paper, we propose DANES, a Deep Neural Network Ensemble Architecture for Social and Textual Context-aware Fake News Detection. DANES comprises a Text Branch for a textual content-based context and a Social Branch for the social context. These two branches are used to create a novel Network Embedding. Preliminary ablation results on 3 real-world datasets, i.e., BuzzFace, Twitter15, and Twitter16, are promising, with an accuracy that outperforms state-of-the-art solutions when employing both social and textual content features.
♻ ☆ Chain of Explanation: New Prompting Method to Generate Higher Quality Natural Language Explanation for Implicit Hate Speech
Recent studies have exploited advanced generative language models to generate Natural Language Explanations (NLE) for why a certain text could be hateful. We propose the Chain of Explanation (CoE) Prompting method, using the target group and retrieved social norms, to generate high-quality NLE for implicit hate speech. Providing accurate target information and high-quality related social norms, we improved the BLUE score from 44.0 to 62.3 for NLE generation. We then evaluate the quality of generated NLE from various automatic metrics and human annotations of informativeness and clarity scores. The correlation analysis between auto-metrics and human perceptions reveals insights into how to select suitable automatic metrics for Natural Language Generation tasks. To showcase a potential application of our proposed CoE method, we demonstrate the f1-score improvements from 0.635 to 0.655 for the implicit hate speech classification task.
♻ ☆ Using novel data and ensemble models to improve automated labeling of Sustainable Development Goals
A number of labeling systems based on text have been proposed to help monitor work on the United Nations (UN) Sustainable Development Goals (SDGs). Here, we present a systematic comparison of systems using a variety of text sources and show that systems differ considerably in their specificity (i.e., true-positive rate) and sensitivity (i.e., true-negative rate), have systematic biases (e.g., are more sensitive to specific SDGs relative to others), and are susceptible to the type and amount of text analyzed. We then show that an ensemble model that pools labeling systems alleviates some of these limitations, exceeding the labeling performance of all currently available systems. We conclude that researchers and policymakers should care about the choice of labeling system and that ensemble methods should be favored when drawing conclusions about the absolute and relative prevalence of work on the SDGs based on automated methods.
♻ ☆ Automatic generation of semantic corpora for improving intent estimation of taxonomy-driven search engines
With the increasing demand of intelligent systems capable of operating in different user contexts (e.g. users on the move) the correct interpretation of the user-need by such systems has become crucial to give a consistent answer to the user query. The most effective techniques which are used to address such task are in the fields of natural language processing and semantic expansion of terms. Such systems are aimed at estimating the actual meaning of input queries, addressing the concepts of the words which are expressed within the user questions. The aim of this paper is to demonstrate which semantic relation impacts the most in semantic expansion-based retrieval systems and to identify the best tradeoff between accuracy and noise introduction when combining such relations. The evaluations are made building a simple natural language processing system capable of querying any taxonomy-driven domain, making use of the combination of different semantic expansions as knowledge resources. The proposed evaluation employs a wide and varied taxonomy as a use-case, exploiting its labels as basis for the expansions. To build the knowledge resources several corpora have been produced and integrated as gazetteers into the NLP infrastructure with the purpose of estimating the pseudo-queries corresponding to the taxonomy labels, considered as the possible intents.
comment: Work in progress
♻ ☆ Towards Answering Open-ended Ethical Quandary Questions
Considerable advancements have been made in various NLP tasks based on the impressive power of large language models (LLMs) and many NLP applications are deployed in our daily lives. In this work, we challenge the capability of LLMs with the new task of Ethical Quandary Generative Question Answering. Ethical quandary questions are more challenging to address because multiple conflicting answers may exist to a single quandary. We explore the current capability of LLMs in providing an answer with a deliberative exchange of different perspectives to an ethical quandary, in the approach of Socratic philosophy, instead of providing a closed answer like an oracle. We propose a model that searches for different ethical principles applicable to the ethical quandary and generates an answer conditioned on the chosen principles through prompt-based few-shot learning. We also discuss the remaining challenges and ethical issues involved in this task and suggest the direction toward developing responsible NLP systems by incorporating human values explicitly.
comment: 16 pages
♻ ☆ NLP as a Lens for Causal Analysis and Perception Mining to Infer Mental Health on Social Media
Interactions among humans on social media often convey intentions behind their actions, yielding a psychological language resource for Mental Health Analysis (MHA) of online users. The success of Computational Intelligence Techniques (CIT) for inferring mental illness from such social media resources points to NLP as a lens for causal analysis and perception mining. However, we argue that more consequential and explainable research is required for optimal impact on clinical psychology practice and personalized mental healthcare. To bridge this gap, we posit two significant dimensions: (1) Causal analysis to illustrate a cause and effect relationship in the user generated text; (2) Perception mining to infer psychological perspectives of social effects on online users intentions. Within the scope of Natural Language Processing (NLP), we further explore critical areas of inquiry associated with these two dimensions, specifically through recent advancements in discourse analysis. This position paper guides the community to explore solutions in this space and advance the state of practice in developing conversational agents for inferring mental health from social media. We advocate for a more explainable approach toward modeling computational psychology problems through the lens of language as we observe an increased number of research contributions in dataset and problem formulation for causal relation extraction and perception enhancements while inferring mental states.
comment: Will revise work
♻ ☆ FADO: Feedback-Aware Double COntrolling Network for Emotional Support Conversation SC
Emotional Support Conversation (ESConv) aims to reduce help-seekers'emotional distress with the supportive strategy and response. It is essential for the supporter to select an appropriate strategy with the feedback of the help-seeker (e.g., emotion change during dialog turns, etc) in ESConv. However, previous methods mainly focus on the dialog history to select the strategy and ignore the help-seeker's feedback, leading to the wrong and user-irrelevant strategy prediction. In addition, these approaches only model the context-to-strategy flow and pay less attention to the strategy-to-context flow that can focus on the strategy-related context for generating the strategy-constrain response. In this paper, we propose a Feedback-Aware Double COntrolling Network (FADO) to make a strategy schedule and generate the supportive response. The core module in FADO consists of a dual-level feedback strategy selector and a double control reader. Specifically, the dual-level feedback strategy selector leverages the turn-level and conversation-level feedback to encourage or penalize strategies. The double control reader constructs the novel strategy-to-context flow for generating the strategy-constrain response. Furthermore, a strategy dictionary is designed to enrich the semantic information of the strategy and improve the quality of strategy-constrain response. Experimental results on ESConv show that the proposed FADO has achieved the state-of-the-art performance in terms of both strategy selection and response generation. Our code is available at https://github/after/reviewing.
comment: Accepted on Knowl. Based Syst. (SCI I)
♻ ☆ REPLUG: Retrieval-Augmented Black-Box Language Models
We introduce REPLUG, a retrieval-augmented language modeling framework that treats the language model (LM) as a black box and augments it with a tuneable retrieval model. Unlike prior retrieval-augmented LMs that train language models with special cross attention mechanisms to encode the retrieved text, REPLUG simply prepends retrieved documents to the input for the frozen black-box LM. This simple design can be easily applied to any existing retrieval and language models. Furthermore, we show that the LM can be used to supervise the retrieval model, which can then find documents that help the LM make better predictions. Our experiments demonstrate that REPLUG with the tuned retriever significantly improves the performance of GPT-3 (175B) on language modeling by 6.3%, as well as the performance of Codex on five-shot MMLU by 5.1%.
♻ ☆ Faithful Chain-of-Thought Reasoning
While Chain-of-Thought (CoT) prompting boosts Language Models' (LM) performance on a gamut of complex reasoning tasks, the generated reasoning chain does not necessarily reflect how the model arrives at the answer (aka. faithfulness). We propose Faithful CoT, a faithful-by-construction framework that decomposes a reasoning task into two stages: Translation (Natural Language query $\rightarrow$ symbolic reasoning chain) and Problem Solving (reasoning chain $\rightarrow$ answer), using an LM and a deterministic solver respectively. We demonstrate the efficacy of our approach on 10 reasoning datasets from 4 diverse domains. It outperforms traditional CoT prompting on 9 out of the 10 datasets, with an average accuracy gain of 4.4 on Math Word Problems, 1.9 on Planning, 4.0 on Multi-hop Question Answering (QA), and 18.1 on Logical Inference, under greedy decoding. Together with self-consistency decoding, we achieve new state-of-the-art few-shot performance on 7 out of the 10 datasets, showing a strong synergy between faithfulness and accuracy.
Computer Vision and Pattern Recognition 74
☆ ADAPT: Action-aware Driving Caption Transformer ICRA2023
End-to-end autonomous driving has great potential in the transportation industry. However, the lack of transparency and interpretability of the automatic decision-making process hinders its industrial adoption in practice. There have been some early attempts to use attention maps or cost volume for better model explainability which is difficult for ordinary passengers to understand. To bridge the gap, we propose an end-to-end transformer-based architecture, ADAPT (Action-aware Driving cAPtion Transformer), which provides user-friendly natural language narrations and reasoning for each decision making step of autonomous vehicular control and action. ADAPT jointly trains both the driving caption task and the vehicular control prediction task, through a shared video representation. Experiments on BDD-X (Berkeley DeepDrive eXplanation) dataset demonstrate state-of-the-art performance of the ADAPT framework on both automatic metrics and human evaluation. To illustrate the feasibility of the proposed framework in real-world applications, we build a novel deployable system that takes raw car videos as input and outputs the action narrations and reasoning in real time. The code, models and data are available at https://github.com/jxbbb/ADAPT.
comment: Accepted to ICRA2023. Code: https://github.com/jxbbb/ADAPT
☆ Stable Target Field for Reduced Variance Score Estimation in Diffusion Models ICLR 2023
Diffusion models generate samples by reversing a fixed forward diffusion process. Despite already providing impressive empirical results, these diffusion models algorithms can be further improved by reducing the variance of the training targets in their denoising score-matching objective. We argue that the source of such variance lies in the handling of intermediate noise-variance scales, where multiple modes in the data affect the direction of reverse paths. We propose to remedy the problem by incorporating a reference batch which we use to calculate weighted conditional scores as more stable training targets. We show that the procedure indeed helps in the challenging intermediate regime by reducing (the trace of) the covariance of training targets. The new stable targets can be seen as trading bias for reduced variance, where the bias vanishes with increasing reference batch size. Empirically, we show that the new objective improves the image quality, stability, and training speed of various popular diffusion models across datasets with both general ODE and SDE solvers. When used in combination with EDM, our method yields a current SOTA FID of 1.90 with 35 network evaluations on the unconditional CIFAR-10 generation task. The code is available at https://github.com/Newbeeer/stf
comment: Accepted by ICLR 2023. Code available at: https://github.com/Newbeeer/stf
☆ Detecting Histologic Glioblastoma Regions of Prognostic Relevance
Glioblastoma is the most common and aggressive malignant adult tumor of the central nervous system, with grim prognosis and heterogeneous morphologic and molecular profiles. Since the adoption of the current standard of care treatment, 18 years ago, there are no substantial prognostic improvements noticed. Accurate prediction of patient overall survival (OS) from clinical histopathology whole slide images (WSI) using advanced computational methods could contribute to optimization of clinical decision making and patient management. Here, we focus on identifying prognostically relevant glioblastoma morphologic patterns on H&E stained WSI. The exact approach capitalizes on the comprehensive WSI curation of apparent artifactual content and on an interpretability mechanism via a weakly supervised attention based multiple instance learning algorithm that further utilizes clustering to constrain the search space. The automatically identified patterns of high diagnostic value are used to classify the WSI as representative of a short or a long survivor. Identifying tumor morphologic patterns associated with short and long OS will allow the clinical neuropathologist to provide additional prognostic information gleaned during microscopic assessment to the treating team, as well as suggest avenues of biological investigation for understanding and potentially treating glioblastoma.
☆ Image-Based Vehicle Classification by Synergizing Features from Supervised and Self-Supervised Learning Paradigms
This paper introduces a novel approach to leverage features learned from both supervised and self-supervised paradigms, to improve image classification tasks, specifically for vehicle classification. Two state-of-the-art self-supervised learning methods, DINO and data2vec, were evaluated and compared for their representation learning of vehicle images. The former contrasts local and global views while the latter uses masked prediction on multi-layered representations. In the latter case, supervised learning is employed to finetune a pretrained YOLOR object detector for detecting vehicle wheels, from which definitive wheel positional features are retrieved. The representations learned from these self-supervised learning methods were combined with the wheel positional features for the vehicle classification task. Particularly, a random wheel masking strategy was utilized to finetune the previously learned representations in harmony with the wheel positional features during the training of the classifier. Our experiments show that the data2vec-distilled representations, which are consistent with our wheel masking strategy, outperformed the DINO counterpart, resulting in a celebrated Top-1 classification accuracy of 97.2% for classifying the 13 vehicle classes defined by the Federal Highway Administration.
comment: 15 pages, 7 figures, 7 tables
☆ Deep Dependency Networks for Multi-Label Classification
We propose a simple approach which combines the strengths of probabilistic graphical models and deep learning architectures for solving the multi-label classification task, focusing specifically on image and video data. First, we show that the performance of previous approaches that combine Markov Random Fields with neural networks can be modestly improved by leveraging more powerful methods such as iterative join graph propagation, integer linear programming, and $\ell_1$ regularization-based structure learning. Then we propose a new modeling framework called deep dependency networks, which augments a dependency network, a model that is easy to train and learns more accurate dependencies but is limited to Gibbs sampling for inference, to the output layer of a neural network. We show that despite its simplicity, jointly learning this new architecture yields significant improvements in performance over the baseline neural network. In particular, our experimental evaluation on three video activity classification datasets: Charades, Textually Annotated Cooking Scenes (TACoS), and Wetlab, and three multi-label image classification datasets: MS-COCO, PASCAL VOC, and NUS-WIDE show that deep dependency networks are almost always superior to pure neural architectures that do not use dependency networks.
☆ Continuous U-Net: Faster, Greater and Noiseless
Image segmentation is a fundamental task in image analysis and clinical practice. The current state-of-the-art techniques are based on U-shape type encoder-decoder networks with skip connections, called U-Net. Despite the powerful performance reported by existing U-Net type networks, they suffer from several major limitations. Issues include the hard coding of the receptive field size, compromising the performance and computational cost, as well as the fact that they do not account for inherent noise in the data. They have problems associated with discrete layers, and do not offer any theoretical underpinning. In this work we introduce continuous U-Net, a novel family of networks for image segmentation. Firstly, continuous U-Net is a continuous deep neural network that introduces new dynamic blocks modelled by second order ordinary differential equations. Secondly, we provide theoretical guarantees for our network demonstrating faster convergence, higher robustness and less sensitivity to noise. Thirdly, we derive qualitative measures to tailor-made segmentation tasks. We demonstrate, through extensive numerical and visual results, that our model outperforms existing U-Net blocks for several medical image segmentation benchmarking datasets.
☆ Transforming CLIP to an Open-vocabulary Video Model via Interpolated Weight Optimization
Contrastive Language-Image Pretraining (CLIP) has demonstrated impressive zero-shot learning abilities for image understanding, yet limited effort has been made to investigate CLIP for zero-shot video recognition. We introduce Open-VCLIP, a simple yet effective approach that transforms CLIP into strong zero-shot video classifiers that can recognize unseen actions and events at test time. Our framework extends CLIP with minimal modifications to model spatial-temporal relationships in videos, making it a specialized video classifier, while striving for generalization. We formally show that training an Open-VCLIP is equivalent to continual learning with zero historical data. To address this problem, we propose Interpolated Weight Optimization, which utilizes the benefit of weight interpolation in both training and test time. We evaluate our method on three popular and challenging action recognition datasets following various zero-shot evaluation protocols and we demonstrate our approach outperforms state-of-the-art methods by clear margins. In particular, we achieve 87.9%, 58.3%, 81.1% zero-shot accuracy on UCF, HMDB and Kinetics-600 respectively, outperforming state-of-the-art methods by 8.3%, 7.8% and 12.2%.
☆ An automated, geometry-based method for the analysis of hippocampal thickness
The hippocampus is one of the most studied neuroanatomical structures due to its involvement in attention, learning, and memory as well as its atrophy in ageing, neurological, and psychiatric diseases. Hippocampal shape changes, however, are complex and cannot be fully characterized by a single summary metric such as hippocampal volume as determined from MR images. In this work, we propose an automated, geometry-based approach for the unfolding, point-wise correspondence, and local analysis of hippocampal shape features such as thickness and curvature. Starting from an automated segmentation of hippocampal subfields, we create a 3D tetrahedral mesh model as well as a 3D intrinsic coordinate system of the hippocampal body. From this coordinate system, we derive local curvature and thickness estimates as well as a 2D sheet for hippocampal unfolding. We evaluate the performance of our algorithm with a series of experiments to quantify neurodegenerative changes in Mild Cognitive Impairment and Alzheimer's disease dementia. We find that hippocampal thickness estimates detect known differences between clinical groups and can determine the location of these effects on the hippocampal sheet. Further, thickness estimates improve classification of clinical groups and cognitively unimpaired controls when added as an additional predictor. Comparable results are obtained with different datasets and segmentation algorithms. Taken together, we replicate canonical findings on hippocampal volume/shape changes in dementia, extend them by gaining insight into their spatial localization on the hippocampal sheet, and provide additional, complementary information beyond traditional measures. We provide a new set of sensitive processing and analysis tools for the analysis of hippocampal geometry that allows comparisons across studies without relying on image registration or requiring manual intervention.
☆ Correspondence-free online human motion retargeting
We present a novel data-driven framework for unsupervised human motion retargeting which animates a target body shape with a source motion. This allows to retarget motions between different characters by animating a target subject with a motion of a source subject. Our method is correspondence-free,~\ie neither spatial correspondences between the source and target shapes nor temporal correspondences between different frames of the source motion are required. Our proposed method directly animates a target shape with arbitrary sequences of humans in motion, possibly captured using 4D acquisition platforms or consumer devices. Our framework takes into account long-term temporal context of $1$ second during retargeting while accounting for surface details. To achieve this, we take inspiration from two lines of existing work: skeletal motion retargeting, which leverages long-term temporal context at the cost of surface detail, and surface-based retargeting, which preserves surface details without considering long-term temporal context. We unify the advantages of these works by combining a learnt skinning field with a skeletal retargeting approach. During inference, our method runs online,~\ie the input can be processed in a serial way, and retargeting is performed in a single forward pass per frame. Experiments show that including long-term temporal context during training improves the method's accuracy both in terms of the retargeted skeletal motion and the detail preservation. Furthermore, our method generalizes well on unobserved motions and body shapes. We demonstrate that the proposed framework achieves state-of-the-art results on two test datasets.
☆ An Out-of-Domain Synapse Detection Challenge for Microwasp Brain Connectomes
The size of image stacks in connectomics studies now reaches the terabyte and often petabyte scales with a great diversity of appearance across brain regions and samples. However, manual annotation of neural structures, e.g., synapses, is time-consuming, which leads to limited training data often smaller than 0.001\% of the test data in size. Domain adaptation and generalization approaches were proposed to address similar issues for natural images, which were less evaluated on connectomics data due to a lack of out-of-domain benchmarks.
☆ A latent space for unsupervised MR image quality control via artifact assessment SP
Image quality control (IQC) can be used in automated magnetic resonance (MR) image analysis to exclude erroneous results caused by poorly acquired or artifact-laden images. Existing IQC methods for MR imaging generally require human effort to craft meaningful features or label large datasets for supervised training. The involvement of human labor can be burdensome and biased, as labeling MR images based on their quality is a subjective task. In this paper, we propose an automatic IQC method that evaluates the extent of artifacts in MR images without supervision. In particular, we design an artifact encoding network that learns representations of artifacts based on contrastive learning. We then use a normalizing flow to estimate the density of learned representations for unsupervised classification. Our experiments on large-scale multi-cohort MR datasets show that the proposed method accurately detects images with high levels of artifacts, which can inform downstream analysis tasks about potentially flawed data.
comment: Accepted at the International Society for Optics and Photonics - Medical Imaging (SPIE-MI) 2023
☆ Uncertainty-Driven Dense Two-View Structure from Motion
This work introduces an effective and practical solution to the dense two-view structure from motion (SfM) problem. One vital question addressed is how to mindfully use per-pixel optical flow correspondence between two frames for accurate pose estimation -- as perfect per-pixel correspondence between two images is difficult, if not impossible, to establish. With the carefully estimated camera pose and predicted per-pixel optical flow correspondences, a dense depth of the scene is computed. Later, an iterative refinement procedure is introduced to further improve optical flow matching confidence, camera pose, and depth, exploiting their inherent dependency in rigid SfM. The fundamental idea presented is to benefit from per-pixel uncertainty in the optical flow estimation and provide robustness to the dense SfM system via an online refinement. Concretely, we introduce a pipeline consisting of (i) an uncertainty-aware dense optical flow estimation approach that provides per-pixel correspondence with their confidence score of matching; (ii) a weighted dense bundle adjustment formulation that depends on optical flow uncertainty and bidirectional optical flow consistency to refine both pose and depth; (iii) a depth estimation network that considers its consistency with the estimated poses and optical flow respecting epipolar constraint. Extensive experiments show that the proposed approach achieves remarkable depth accuracy and state-of-the-art camera pose results superseding SuperPoint and SuperGlue accuracy when tested on benchmark datasets such as DeMoN, YFCC100M, and ScanNet.
comment: Accepted for publication at IEEE Robotics and Automation Letters (RA-L) 2023
☆ Synthesis-based Imaging-Differentiation Representation Learning for Multi-Sequence 3D/4D MRI
Multi-sequence MRIs can be necessary for reliable diagnosis in clinical practice due to the complimentary information within sequences. However, redundant information exists across sequences, which interferes with mining efficient representations by modern machine learning or deep learning models. To handle various clinical scenarios, we propose a sequence-to-sequence generation framework (Seq2Seq) for imaging-differentiation representation learning. In this study, not only do we propose arbitrary 3D/4D sequence generation within one model to generate any specified target sequence, but also we are able to rank the importance of each sequence based on a new metric estimating the difficulty of a sequence being generated. Furthermore, we also exploit the generation inability of the model to extract regions that contain unique information for each sequence. We conduct extensive experiments using three datasets including a toy dataset of 20,000 simulated subjects, a brain MRI dataset of 1,251 subjects, and a breast MRI dataset of 2,101 subjects, to demonstrate that (1) our proposed Seq2Seq is efficient and lightweight for complex clinical datasets and can achieve excellent image quality; (2) top-ranking sequences can be used to replace complete sequences with non-inferior performance; (3) combining MRI with our imaging-differentiation map leads to better performance in clinical tasks such as glioblastoma MGMT promoter methylation status prediction and breast cancer pathological complete response status prediction. Our code is available at https://github.com/fiy2W/mri_seq2seq.
☆ Towards Implementing Energy-aware Data-driven Intelligence for Smart Health Applications on Mobile Platforms
Recent breakthrough technological progressions of powerful mobile computing resources such as low-cost mobile GPUs along with cutting-edge, open-source software architectures have enabled high-performance deep learning on mobile platforms. These advancements have revolutionized the capabilities of today's mobile applications in different dimensions to perform data-driven intelligence locally, particularly for smart health applications. Unlike traditional machine learning (ML) architectures, modern on-device deep learning frameworks are proficient in utilizing computing resources in mobile platforms seamlessly, in terms of producing highly accurate results in less inference time. However, on the flip side, energy resources in a mobile device are typically limited. Hence, whenever a complex Deep Neural Network (DNN) architecture is fed into the on-device deep learning framework, while it achieves high prediction accuracy (and performance), it also urges huge energy demands during the runtime. Therefore, managing these resources efficiently within the spectrum of performance and energy efficiency is the newest challenge for any mobile application featuring data-driven intelligence beyond experimental evaluations. In this paper, first, we provide a timely review of recent advancements in on-device deep learning while empirically evaluating the performance metrics of current state-of-the-art ML architectures and conventional ML approaches with the emphasis given on energy characteristics by deploying them on a smart health application. With that, we are introducing a new framework through an energy-aware, adaptive model comprehension and realization (EAMCR) approach that can be utilized to make more robust and efficient inference decisions based on the available computing/energy resources in the mobile device during the runtime.
☆ Exploring Semantic Perturbations on Grover
With news and information being as easy to access as they currently are, it is more important than ever to ensure that people are not mislead by what they read. Recently, the rise of neural fake news (AI-generated fake news) and its demonstrated effectiveness at fooling humans has prompted the development of models to detect it. One such model is the Grover model, which can both detect neural fake news to prevent it, and generate it to demonstrate how a model could be misused to fool human readers. In this work we explore the Grover model's fake news detection capabilities by performing targeted attacks through perturbations on input news articles. Through this we test Grover's resilience to these adversarial attacks and expose some potential vulnerabilities which should be addressed in further iterations to ensure it can detect all types of fake news accurately.
comment: 15 pages, 12 figures, 1 table, capstone research in machine learning
☆ Tracking People in Highly Dynamic Industrial Environments
To date, the majority of positioning systems have been designed to operate within environments that have long-term stable macro-structure with potential small-scale dynamics. These assumptions allow the existing positioning systems to produce and utilize stable maps. However, in highly dynamic industrial settings these assumptions are no longer valid and the task of tracking people is more challenging due to the rapid large-scale changes in structure. In this paper we propose a novel positioning system for tracking people in highly dynamic industrial environments, such as construction sites. The proposed system leverages the existing CCTV camera infrastructure found in many industrial settings along with radio and inertial sensors within each worker's mobile phone to accurately track multiple people. This multi-target multi-sensor tracking framework also allows our system to use cross-modality training in order to deal with the environment dynamics. In particular, we show how our system uses cross-modality training in order to automatically keep track environmental changes (i.e. new walls) by utilizing occlusion maps. In addition, we show how these maps can be used in conjunction with social forces to accurately predict human motion and increase the tracking accuracy. We have conducted extensive real-world experiments in a construction site showing significant accuracy improvement via cross-modality training and the use of social forces.
☆ Learning Prototype Classifiers for Long-Tailed Recognition
The problem of long-tailed recognition (LTR) has received attention in recent years due to the fundamental power-law distribution of objects in the real-world. Most recent works in LTR use softmax classifiers that have a tendency to correlate classifier norm with the amount of training data for a given class. On the other hand, Prototype classifiers do not suffer from this shortcoming and can deliver promising results simply using Nearest-Class-Mean (NCM), a special case where prototypes are empirical centroids. However, the potential of Prototype classifiers as an alternative to softmax in LTR is relatively underexplored. In this work, we propose Prototype classifiers, which jointly learn prototypes that minimize average cross-entropy loss based on probability scores from distances to prototypes. We theoretically analyze the properties of Euclidean distance based prototype classifiers that leads to stable gradient-based optimization which is robust to outliers. We further enhance Prototype classifiers by learning channel-dependent temperature parameters to enable independent distance scales along each channel. Our analysis shows that prototypes learned by Prototype classifiers are better separated than empirical centroids. Results on four long-tailed recognition benchmarks show that Prototype classifier outperforms or is comparable to the state-of-the-art methods.
comment: 11 pages, 4 figures and 9 tables
☆ Do I Have Your Attention: A Large Scale Engagement Prediction Dataset and Baselines
The degree of concentration, enthusiasm, optimism, and passion displayed by individual(s) while interacting with a machine is referred to as `user engagement'. Engagement comprises of behavioural, cognitive, and affect related cues. To create engagement predictions systems, which can work in real-world conditions it is quintessential to learn from rich diverse datasets. To this end, a large scale multi-faceted engagement in the wild dataset is proposed. 31 hours duration data of 127 participants representing different illumination conditions is recorded. Thorough experiments are performed exploring applicability of different features action units, eye gaze and head pose and transformers. To further validate the rich nature of the dataset, evaluation is also performed on the EngageWild dataset. The experiments show the usefulness of the proposed dataset. The code, models and dataset will be made publicly available.
☆ mPLUG-2: A Modularized Multi-modal Foundation Model Across Text, Image and Video
Recent years have witnessed a big convergence of language, vision, and multi-modal pretraining. In this work, we present mPLUG-2, a new unified paradigm with modularized design for multi-modal pretraining, which can benefit from modality collaboration while addressing the problem of modality entanglement. In contrast to predominant paradigms of solely relying on sequence-to-sequence generation or encoder-based instance discrimination, mPLUG-2 introduces a multi-module composition network by sharing common universal modules for modality collaboration and disentangling different modality modules to deal with modality entanglement. It is flexible to select different modules for different understanding and generation tasks across all modalities including text, image, and video. Empirical study shows that mPLUG-2 achieves state-of-the-art or competitive results on a broad range of over 30 downstream tasks, spanning multi-modal tasks of image-text and video-text understanding and generation, and uni-modal tasks of text-only, image-only, and video-only understanding. Notably, mPLUG-2 shows new state-of-the-art results of 48.0 top-1 accuracy and 80.3 CIDEr on the challenging MSRVTT video QA and video caption tasks with a far smaller model size and data scale. It also demonstrates strong zero-shot transferability on vision-language and video-language tasks. Code and models will be released in https://github.com/alibaba/AliceMind.
☆ PresSim: An End-to-end Framework for Dynamic Ground Pressure Profile Generation from Monocular Videos Using Physics-based 3D Simulation
Ground pressure exerted by the human body is a valuable source of information for human activity recognition (HAR) in unobtrusive pervasive sensing. While data collection from pressure sensors to develop HAR solutions requires significant resources and effort, we present a novel end-to-end framework, PresSim, to synthesize sensor data from videos of human activities to reduce such effort significantly. PresSim adopts a 3-stage process: first, extract the 3D activity information from videos with computer vision architectures; then simulate the floor mesh deformation profiles based on the 3D activity information and gravity-included physics simulation; lastly, generate the simulated pressure sensor data with deep learning models. We explored two approaches for the 3D activity information: inverse kinematics with mesh re-targeting, and volumetric pose and shape estimation. We validated PresSim with an experimental setup with a monocular camera to provide input and a pressure-sensing fitness mat (80x28 spatial resolution) to provide the sensor ground truth, where nine participants performed a set of predefined yoga sequences.
comment: Percom2023 workshop(UMUM2023)
☆ EfficientRep:An Efficient Repvgg-style ConvNets with Hardware-aware Neural Network Design
We present a hardware-efficient architecture of convolutional neural network, which has a repvgg-like architecture. Flops or parameters are traditional metrics to evaluate the efficiency of networks which are not sensitive to hardware including computing ability and memory bandwidth. Thus, how to design a neural network to efficiently use the computing ability and memory bandwidth of hardware is a critical problem. This paper proposes a method how to design hardware-aware neural network. Based on this method, we designed EfficientRep series convolutional networks, which are high-computation hardware(e.g. GPU) friendly and applied in YOLOv6 object detection framework. YOLOv6 has published YOLOv6N/YOLOv6S/YOLOv6M/YOLOv6L models in v1 and v2 versions.
☆ Alphazzle: Jigsaw Puzzle Solver with Deep Monte-Carlo Tree Search
Solving jigsaw puzzles requires to grasp the visual features of a sequence of patches and to explore efficiently a solution space that grows exponentially with the sequence length. Therefore, visual deep reinforcement learning (DRL) should answer this problem more efficiently than optimization solvers coupled with neural networks. Based on this assumption, we introduce Alphazzle, a reassembly algorithm based on single-player Monte Carlo Tree Search (MCTS). A major difference with DRL algorithms lies in the unavailability of game reward for MCTS, and we show how to estimate it from the visual input with neural networks. This constraint is induced by the puzzle-solving task and dramatically adds to the task complexity (and interest!). We perform an in-deep ablation study that shows the importance of MCTS and the neural networks working together. We achieve excellent results and get exciting insights into the combination of DRL and visual feature learning.
☆ Test-Time Amendment with a Coarse Classifier for Fine-Grained Classification
We investigate the problem of reducing mistake severity for fine-grained classification. Fine-grained classification can be challenging, mainly due to the requirement of knowledge or domain expertise for accurate annotation. However, humans are particularly adept at performing coarse classification as it requires relatively low levels of expertise. To this end, we present a novel approach for Post-Hoc Correction called Hierarchical Ensembles (HiE) that utilizes label hierarchy to improve the performance of fine-grained classification at test-time using the coarse-grained predictions. By only requiring the parents of leaf nodes, our method significantly reduces avg. mistake severity while improving top-1 accuracy on the iNaturalist-19 and tieredImageNet-H datasets, achieving a new state-of-the-art on both benchmarks. We also investigate the efficacy of our approach in the semi-supervised setting. Our approach brings notable gains in top-1 accuracy while significantly decreasing the severity of mistakes as training data decreases for the fine-grained classes. The simplicity and post-hoc nature of HiE render it practical to be used with any off-the-shelf trained model to improve its predictions further.
comment: 8 pages, 2 figures, 3 tables
☆ A Flexible Framework for Virtual Omnidirectional Vision to Improve Operator Situation Awareness
During teleoperation of a mobile robot, providing good operator situation awareness is a major concern as a single mistake can lead to mission failure. Camera streams are widely used for teleoperation but offer limited field-of-view. In this paper, we present a flexible framework for virtual projections to increase situation awareness based on a novel method to fuse multiple cameras mounted anywhere on the robot. Moreover, we propose a complementary approach to improve scene understanding by fusing camera images and geometric 3D Lidar data to obtain a colorized point cloud. The implementation on a compact omnidirectional camera reduces system complexity considerably and solves multiple use-cases on a much smaller footprint compared to traditional approaches such as actuated pan-tilt units. Finally, we demonstrate the generality of the approach by application to the multi-camera system of the Boston Dynamics Spot. The software implementation is available as open-source ROS packages on the project page https://tu-darmstadt-ros-pkg.github.io/omnidirectional_vision.
comment: Accepted to European Conference on Mobile Robots (ECMR) 2021. Video link: https://youtu.be/7pocpdsMxOM Project page: https://tu-darmstadt-ros-pkg.github.io/omnidirectional_vision
☆ Towards Label-Efficient Incremental Learning: A Survey
The current dominant paradigm when building a machine learning model is to iterate over a dataset over and over until convergence. Such an approach is non-incremental, as it assumes access to all images of all categories at once. However, for many applications, non-incremental learning is unrealistic. To that end, researchers study incremental learning, where a learner is required to adapt to an incoming stream of data with a varying distribution while preventing forgetting of past knowledge. Significant progress has been made, however, the vast majority of works focus on the fully supervised setting, making these algorithms label-hungry thus limiting their real-life deployment. To that end, in this paper, we make the first attempt to survey recently growing interest in label-efficient incremental learning. We identify three subdivisions, namely semi-, few-shot- and self-supervised learning to reduce labeling efforts. Finally, we identify novel directions that can further enhance label-efficiency and improve incremental learning scalability. Project website: {https://github.com/kilickaya/label-efficient-il.
☆ iPAL: A Machine Learning Based Smart Healthcare Framework For Automatic Diagnosis Of Attention Deficit/Hyperactivity Disorder (ADHD)
ADHD is a prevalent disorder among the younger population. Standard evaluation techniques currently use evaluation forms, interviews with the patient, and more. However, its symptoms are similar to those of many other disorders like depression, conduct disorder, and oppositional defiant disorder, and these current diagnosis techniques are not very effective. Thus, a sophisticated computing model holds the potential to provide a promising diagnosis solution to this problem. This work attempts to explore methods to diagnose ADHD using combinations of multiple established machine learning techniques like neural networks and SVM models on the ADHD200 dataset and explore the field of neuroscience. In this work, multiclass classification is performed on phenotypic data using an SVM model. The better results have been analyzed on the phenotypic data compared to other supervised learning techniques like Logistic regression, KNN, AdaBoost, etc. In addition, neural networks have been implemented on functional connectivity from the MRI data of a sample of 40 subjects provided to achieve high accuracy without prior knowledge of neuroscience. It is combined with the phenotypic classifier using the ensemble technique to get a binary classifier. It is further trained and tested on 400 out of 824 subjects from the ADHD200 data set and achieved an accuracy of 92.5% for binary classification The training and testing accuracy has been achieved upto 99% using ensemble classifier.
☆ Development of Real-time Rendering Technology for High-Precision Models in Autonomous Driving
Our autonomous driving simulation lab produces a high-precision 3D model simulating the parking lot. However, the current model still has poor rendering quality in some aspects. In this work, we develop a system to improve the rendering of the model and evaluate the quality of the rendered model.
comment: 3 pages, 6 figures
☆ Multispectral Pedestrian Detection via Reference Box Constrained Cross Attention and Modality Balanced Optimization
Multispectral pedestrian detection is an important task for many around-the-clock applications, since the visible and thermal modalities can provide complementary information especially under low light conditions. To reduce the influence of hand-designed components in available multispectral pedestrian detectors, we propose a MultiSpectral pedestrian DEtection TRansformer (MS-DETR), which extends deformable DETR to multi-modal paradigm. In order to facilitate the multi-modal learning process, a Reference box Constrained Cross-Attention (RCCA) module is firstly introduced to the multi-modal Transformer decoder, which takes fusion branch together with the reference boxes as intermediaries to enable the interaction of visible and thermal modalities. To further balance the contribution of different modalities, we design a modality-balanced optimization strategy, which aligns the slots of decoders by adaptively adjusting the instance-level weight of three branches. Our end-to-end MS-DETR shows superior performance on the challenging KAIST and CVC-14 benchmark datasets.
☆ Learning Generalized Zero-Shot Learners for Open-Domain Image Geolocalization
Image geolocalization is the challenging task of predicting the geographic coordinates of origin for a given photo. It is an unsolved problem relying on the ability to combine visual clues with general knowledge about the world to make accurate predictions across geographies. We present $\href{https://huggingface.co/geolocal/StreetCLIP}{\text{StreetCLIP}}$, a robust, publicly available foundation model not only achieving state-of-the-art performance on multiple open-domain image geolocalization benchmarks but also doing so in a zero-shot setting, outperforming supervised models trained on more than 4 million images. Our method introduces a meta-learning approach for generalized zero-shot learning by pretraining CLIP from synthetic captions, grounding CLIP in a domain of choice. We show that our method effectively transfers CLIP's generalized zero-shot capabilities to the domain of image geolocalization, improving in-domain generalized zero-shot performance without finetuning StreetCLIP on a fixed set of classes.
☆ Compositional Prompt Tuning with Motion Cues for Open-vocabulary Video Relation Detection ICLR 2023
Prompt tuning with large-scale pretrained vision-language models empowers open-vocabulary predictions trained on limited base categories, e.g., object classification and detection. In this paper, we propose compositional prompt tuning with motion cues: an extended prompt tuning paradigm for compositional predictions of video data. In particular, we present Relation Prompt (RePro) for Open-vocabulary Video Visual Relation Detection (Open-VidVRD), where conventional prompt tuning is easily biased to certain subject-object combinations and motion patterns. To this end, RePro addresses the two technical challenges of Open-VidVRD: 1) the prompt tokens should respect the two different semantic roles of subject and object, and 2) the tuning should account for the diverse spatio-temporal motion patterns of the subject-object compositions. Without bells and whistles, our RePro achieves a new state-of-the-art performance on two VidVRD benchmarks of not only the base training object and predicate categories, but also the unseen ones. Extensive ablations also demonstrate the effectiveness of the proposed compositional and multi-mode design of prompts. Code is available at https://github.com/Dawn-LX/OpenVoc-VidVRD.
comment: accepted by ICLR 2023
☆ SPIDE: A Purely Spike-based Method for Training Feedback Spiking Neural Networks
Spiking neural networks (SNNs) with event-based computation are promising brain-inspired models for energy-efficient applications on neuromorphic hardware. However, most supervised SNN training methods, such as conversion from artificial neural networks or direct training with surrogate gradients, require complex computation rather than spike-based operations of spiking neurons during training. In this paper, we study spike-based implicit differentiation on the equilibrium state (SPIDE) that extends the recently proposed training method, implicit differentiation on the equilibrium state (IDE), for supervised learning with purely spike-based computation, which demonstrates the potential for energy-efficient training of SNNs. Specifically, we introduce ternary spiking neuron couples and prove that implicit differentiation can be solved by spikes based on this design, so the whole training procedure, including both forward and backward passes, is made as event-driven spike computation, and weights are updated locally with two-stage average firing rates. Then we propose to modify the reset membrane potential to reduce the approximation error of spikes. With these key components, we can train SNNs with flexible structures in a small number of time steps and with firing sparsity during training, and the theoretical estimation of energy costs demonstrates the potential for high efficiency. Meanwhile, experiments show that even with these constraints, our trained models can still achieve competitive results on MNIST, CIFAR-10, CIFAR-100, and CIFAR10-DVS. Our code is available at https://github.com/pkuxmq/SPIDE-FSNN.
comment: Accepted by Neural Networks
☆ The Past, Current, and Future of Neonatal Intensive Care Units with Artificial Intelligence
Artificial intelligence (AI), specifically a branch of AI called deep learning (DL), has proven revolutionary developments in almost all fields, from computer vision to health sciences, and its effects in medicine have changed clinical applications significantly. Although some sub-fields of medicine such as pediatrics have been relatively slow in receiving critical benefits of AI, related research in pediatrics started to be accumulated to a significant level too. Hence, in this paper, we review recently developed machine learning and deep learning based systems for neonatology applications. We systematically evaluate the role of AI in neonatology applications, define the methodologies, including algorithmic developments, and describe the remaining challenges in neonatal diseases. To date, survival analysis, neuroimaging, EEG, pattern analysis of vital parameters, and retinopathy of prematurity diagnosis with AI have been the main focus in neonatology. We have categorically summarized 96 research articles, from 1996 to 2022, and discussed their pros and cons, respectively. We also discuss possible directions for new AI models and the future of neonatology with the rising power of AI, suggesting roadmaps for integration of AI into neonatal intensive care units.
comment: 58 pages, review article
☆ Human Fall Detection- Multimodality Approach
Falls have become more frequent in recent years, which has been harmful for senior citizens.Therefore detecting falls have become important and several data sets and machine learning model have been introduced related to fall detection. In this project report, a human fall detection method is proposed using a multi modality approach. We used the UP-FALL detection data set which is collected by dozens of volunteers using different sensors and two cameras. We use wrist sensor with acclerometer data keeping labels to binary classification, namely fall and no fall from the data set.We used fusion of camera and sensor data to increase performance. The experimental results shows that using only wrist data as compared to multi sensor for binary classification did not impact the model prediction performance for fall detection.
☆ Efficient Scopeformer: Towards Scalable and Rich Feature Extraction for Intracranial Hemorrhage Detection
The quality and richness of feature maps extracted by convolution neural networks (CNNs) and vision Transformers (ViTs) directly relate to the robust model performance. In medical computer vision, these information-rich features are crucial for detecting rare cases within large datasets. This work presents the "Scopeformer," a novel multi-CNN-ViT model for intracranial hemorrhage classification in computed tomography (CT) images. The Scopeformer architecture is scalable and modular, which allows utilizing various CNN architectures as the backbone with diversified output features and pre-training strategies. We propose effective feature projection methods to reduce redundancies among CNN-generated features and to control the input size of ViTs. Extensive experiments with various Scopeformer models show that the model performance is proportional to the number of convolutional blocks employed in the feature extractor. Using multiple strategies, including diversifying the pre-training paradigms for CNNs, different pre-training datasets, and style transfer techniques, we demonstrate an overall improvement in the model performance at various computational budgets. Later, we propose smaller compute-efficient Scopeformer versions with three different types of input and output ViT configurations. Efficient Scopeformers use four different pre-trained CNN architectures as feature extractors to increase feature richness. Our best Efficient Scopeformer model achieved an accuracy of 96.94\% and a weighted logarithmic loss of 0.083 with an eight times reduction in the number of trainable parameters compared to the base Scopeformer. Another version of the Efficient Scopeformer model further reduced the parameter space by almost 17 times with negligible performance reduction. Hybrid CNNs and ViTs might provide the desired feature richness for developing accurate medical computer vision models
☆ QCRS: Improve Randomized Smoothing using Quasi-Concave Optimization
Randomized smoothing is currently the state-of-the-art method that provides certified robustness for deep neural networks. However, it often cannot achieve an adequate certified region on real-world datasets. One way to obtain a larger certified region is to use an input-specific algorithm instead of using a fixed Gaussian filter for all data points. Several methods based on this idea have been proposed, but they either suffer from high computational costs or gain marginal improvement in certified radius. In this work, we show that by exploiting the quasiconvex problem structure, we can find the optimal certified radii for most data points with slight computational overhead. This observation leads to an efficient and effective input-specific randomized smoothing algorithm. We conduct extensive experiments and empirical analysis on Cifar10 and ImageNet. The results show that the proposed method significantly enhances the certified radii with low computational overhead.
comment: 10 pages
☆ Density peak clustering using tensor network
Tensor networks, which have been traditionally used to simulate many-body physics, have recently gained significant attention in the field of machine learning due to their powerful representation capabilities. In this work, we propose a density-based clustering algorithm inspired by tensor networks. We encode classical data into tensor network states on an extended Hilbert space and train the tensor network states to capture the features of the clusters. Here, we define density and related concepts in terms of fidelity, rather than using a classical distance measure. We evaluate the performance of our algorithm on six synthetic data sets, four real world data sets, and three commonly used computer vision data sets. The results demonstrate that our method provides state-of-the-art performance on several synthetic data sets and real world data sets, even when the number of clusters is unknown. Additionally, our algorithm performs competitively with state-of-the-art algorithms on the MNIST, USPS, and Fashion-MNIST image data sets. These findings reveal the great potential of tensor networks for machine learning applications.
☆ Neural Wavelet-domain Diffusion for 3D Shape Generation, Inversion, and Manipulation
This paper presents a new approach for 3D shape generation, inversion, and manipulation, through a direct generative modeling on a continuous implicit representation in wavelet domain. Specifically, we propose a compact wavelet representation with a pair of coarse and detail coefficient volumes to implicitly represent 3D shapes via truncated signed distance functions and multi-scale biorthogonal wavelets. Then, we design a pair of neural networks: a diffusion-based generator to produce diverse shapes in the form of the coarse coefficient volumes and a detail predictor to produce compatible detail coefficient volumes for introducing fine structures and details. Further, we may jointly train an encoder network to learn a latent space for inverting shapes, allowing us to enable a rich variety of whole-shape and region-aware shape manipulations. Both quantitative and qualitative experimental results manifest the compelling shape generation, inversion, and manipulation capabilities of our approach over the state-of-the-art methods.
comment: arXiv admin note: substantial text overlap with arXiv:2209.08725
☆ Stable Attribute Group Editing for Reliable Few-shot Image Generation
Few-shot image generation aims to generate data of an unseen category based on only a few samples. Apart from basic content generation, a bunch of downstream applications hopefully benefit from this task, such as low-data detection and few-shot classification. To achieve this goal, the generated images should guarantee category retention for classification beyond the visual quality and diversity. In our preliminary work, we present an ``editing-based'' framework Attribute Group Editing (AGE) for reliable few-shot image generation, which largely improves the generation performance. Nevertheless, AGE's performance on downstream classification is not as satisfactory as expected. This paper investigates the class inconsistency problem and proposes Stable Attribute Group Editing (SAGE) for more stable class-relevant image generation. SAGE takes use of all given few-shot images and estimates a class center embedding based on the category-relevant attribute dictionary. Meanwhile, according to the projection weights on the category-relevant attribute dictionary, we can select category-irrelevant attributes from the similar seen categories. Consequently, SAGE injects the whole distribution of the novel class into StyleGAN's latent space, thus largely remains the category retention and stability of the generated images. Going one step further, we find that class inconsistency is a common problem in GAN-generated images for downstream classification. Even though the generated images look photo-realistic and requires no category-relevant editing, they are usually of limited help for downstream classification. We systematically discuss this issue from both the generative model and classification model perspectives, and propose to boost the downstream classification performance of SAGE by enhancing the pixel and frequency components.
☆ Program Generation from Diverse Video Demonstrations
The ability to use inductive reasoning to extract general rules from multiple observations is a vital indicator of intelligence. As humans, we use this ability to not only interpret the world around us, but also to predict the outcomes of the various interactions we experience. Generalising over multiple observations is a task that has historically presented difficulties for machines to grasp, especially when requiring computer vision. In this paper, we propose a model that can extract general rules from video demonstrations by simultaneously performing summarisation and translation. Our approach differs from prior works by framing the problem as a multi-sequence-to-sequence task, wherein summarisation is learnt by the model. This allows our model to utilise edge cases that would otherwise be suppressed or discarded by traditional summarisation techniques. Additionally, we show that our approach can handle noisy specifications without the need for additional filtering methods. We evaluate our model by synthesising programs from video demonstrations in the Vizdoom environment achieving state-of-the-art results with a relative increase of 11.75% program accuracy on prior works
☆ Detection of Tomato Ripening Stages using Yolov3-tiny
One of the most important agricultural products in Mexico is the tomato (Solanum lycopersicum), which occupies the 4th place national most produced product . Therefore, it is necessary to improve its production, building automatic detection system that detect, classify an keep tacks of the fruits is one way to archieve it. So, in this paper, we address the design of a computer vision system to detect tomatoes at different ripening stages. To solve the problem, we use a neural network-based model for tomato classification and detection. Specifically, we use the YOLOv3-tiny model because it is one of the lightest current deep neural networks. To train it, we perform two grid searches testing several combinations of hyperparameters. Our experiments showed an f1-score of 90.0% in the localization and classification of ripening stages in a custom dataset.
☆ Continual Segment: Towards a Single, Unified and Accessible Continual Segmentation Model of 143 Whole-body Organs in CT Scans
Deep learning empowers the mainstream medical image segmentation methods. Nevertheless current deep segmentation approaches are not capable of efficiently and effectively adapting and updating the trained models when new incremental segmentation classes (along with new training datasets or not) are required to be added. In real clinical environment, it can be preferred that segmentation models could be dynamically extended to segment new organs/tumors without the (re-)access to previous training datasets due to obstacles of patient privacy and data storage. This process can be viewed as a continual semantic segmentation (CSS) problem, being understudied for multi-organ segmentation. In this work, we propose a new architectural CSS learning framework to learn a single deep segmentation model for segmenting a total of 143 whole-body organs. Using the encoder/decoder network structure, we demonstrate that a continually-trained then frozen encoder coupled with incrementally-added decoders can extract and preserve sufficiently representative image features for new classes to be subsequently and validly segmented. To maintain a single network model complexity, we trim each decoder progressively using neural architecture search and teacher-student based knowledge distillation. To incorporate with both healthy and pathological organs appearing in different datasets, a novel anomaly-aware and confidence learning module is proposed to merge the overlapped organ predictions, originated from different decoders. Trained and validated on 3D CT scans of 2500+ patients from four datasets, our single network can segment total 143 whole-body organs with very high accuracy, closely reaching the upper bound performance level by training four separate segmentation models (i.e., one model per dataset/task).
☆ SkinCon: A skin disease dataset densely annotated by domain experts for fine-grained model debugging and analysis NeurIPS 2022
For the deployment of artificial intelligence (AI) in high-risk settings, such as healthcare, methods that provide interpretability/explainability or allow fine-grained error analysis are critical. Many recent methods for interpretability/explainability and fine-grained error analysis use concepts, which are meta-labels that are semantically meaningful to humans. However, there are only a few datasets that include concept-level meta-labels and most of these meta-labels are relevant for natural images that do not require domain expertise. Densely annotated datasets in medicine focused on meta-labels that are relevant to a single disease such as melanoma. In dermatology, skin disease is described using an established clinical lexicon that allows clinicians to describe physical exam findings to one another. To provide a medical dataset densely annotated by domain experts with annotations useful across multiple disease processes, we developed SkinCon: a skin disease dataset densely annotated by dermatologists. SkinCon includes 3230 images from the Fitzpatrick 17k dataset densely annotated with 48 clinical concepts, 22 of which have at least 50 images representing the concept. The concepts used were chosen by two dermatologists considering the clinical descriptor terms used to describe skin lesions. Examples include "plaque", "scale", and "erosion". The same concepts were also used to label 656 skin disease images from the Diverse Dermatology Images dataset, providing an additional external dataset with diverse skin tone representations. We review the potential applications for the SkinCon dataset, such as probing models, concept-based explanations, and concept bottlenecks. Furthermore, we use SkinCon to demonstrate two of these use cases: debugging mistakes of an existing dermatology AI model with concepts and developing interpretable models with post-hoc concept bottleneck models.
comment: NeurIPS 2022 Datasets and Benchmarks Track
♻ ☆ Learning Vortex Dynamics for Fluid Inference and Prediction ICLR 2023
We propose a novel differentiable vortex particle (DVP) method to infer and predict fluid dynamics from a single video. Lying at its core is a particle-based latent space to encapsulate the hidden, Lagrangian vortical evolution underpinning the observable, Eulerian flow phenomena. Our differentiable vortex particles are coupled with a learnable, vortex-to-velocity dynamics mapping to effectively capture the complex flow features in a physically-constrained, low-dimensional space. This representation facilitates the learning of a fluid simulator tailored to the input video that can deliver robust, long-term future predictions. The value of our method is twofold: first, our learned simulator enables the inference of hidden physics quantities (e.g., velocity field) purely from visual observation; secondly, it also supports future prediction, constructing the input video's sequel along with its future dynamics evolution. We compare our method with a range of existing methods on both synthetic and real-world videos, demonstrating improved reconstruction quality, visual plausibility, and physical integrity.
comment: ICLR 2023, project webpage: https://yitongdeng.github.io/vortex_learning_webpage/
♻ ☆ Dynamic Linear Transformer for 3D Biomedical Image Segmentation
Transformer-based neural networks have surpassed promising performance on many biomedical image segmentation tasks due to a better global information modeling from the self-attention mechanism. However, most methods are still designed for 2D medical images while ignoring the essential 3D volume information. The main challenge for 3D transformer-based segmentation methods is the quadratic complexity introduced by the self-attention mechanism \cite{vaswani2017attention}. In this paper, we propose a novel transformer architecture for 3D medical image segmentation using an encoder-decoder style architecture with linear complexity. Furthermore, we newly introduce a dynamic token concept to further reduce the token numbers for self-attention calculation. Taking advantage of the global information modeling, we provide uncertainty maps from different hierarchy stages. We evaluate this method on multiple challenging CT pancreas segmentation datasets. Our promising results show that our novel 3D Transformer-based segmentor could provide promising highly feasible segmentation performance and accurate uncertainty quantification using single annotation. Code is available https://github.com/freshman97/LinTransUNet.
comment: 8 Pages
♻ ☆ 3DShape2VecSet: A 3D Shape Representation for Neural Fields and Generative Diffusion Models
We introduce 3DShape2VecSet, a novel shape representation for neural fields designed for generative diffusion models. Our shape representation can encode 3D shapes given as surface models or point clouds, and represents them as neural fields. The concept of neural fields has previously been combined with a global latent vector, a regular grid of latent vectors, or an irregular grid of latent vectors. Our new representation encodes neural fields on top of a set of vectors. We draw from multiple concepts, such as the radial basis function representation and the cross attention and self-attention function, to design a learnable representation that is especially suitable for processing with transformers. Our results show improved performance in 3D shape encoding and 3D shape generative modeling tasks. We demonstrate a wide variety of generative applications: unconditioned generation, category-conditioned generation, text-conditioned generation, point-cloud completion, and image-conditioned generation.
comment: Project demo: https://youtu.be/KKQsQccpBFk
♻ ☆ Nucleus Segmentation and Analysis in Breast Cancer with the MIScnn Framework
The NuCLS dataset contains over 220.000 annotations of cell nuclei in breast cancers. We show how to use these data to create a multi-rater model with the MIScnn Framework to automate the analysis of cell nuclei. For the model creation, we use the widespread U-Net approach embedded in a pipeline. This pipeline provides besides the high performance convolution neural network, several preprocessor techniques and a extended data exploration. The final model is tested in the evaluation phase using a wide variety of metrics with a subsequent visualization. Finally, the results are compared and interpreted with the results of the NuCLS study. As an outlook, indications are given which are important for the future development of models in the context of cell nuclei.
♻ ☆ Sci-Net: Scale Invariant Model for Buildings Segmentation from Aerial Imagery
Buildings' segmentation is a fundamental task in the field of earth observation and aerial imagery analysis. Most existing deep learning-based methods in the literature can be applied to a fixed or narrow-range spatial resolution imagery. In practical scenarios, users deal with a broad spectrum of image resolutions. Thus, a given aerial image often needs to be re-sampled to match the spatial resolution of the dataset used to train the deep learning model, which results in a degradation in segmentation performance. To overcome this challenge, we propose, in this manuscript, Scale-invariant Neural Network (Sci-Net) architecture that segments buildings from wide-range spatial resolution aerial images. Specifically, our approach leverages UNet hierarchical representation and Dense Atrous Spatial Pyramid Pooling to extract fine-grained multi-scale representations. Sci-Net significantly outperforms state of the art models on the Open Cities AI and the Multi-Scale Building datasets with a steady improvement margin across different spatial resolutions.
♻ ☆ Image Masking for Robust Self-Supervised Monocular Depth Estimation ICRA
Self-supervised monocular depth estimation is a salient task for 3D scene understanding. Learned jointly with monocular ego-motion estimation, several methods have been proposed to predict accurate pixel-wise depth without using labeled data. Nevertheless, these methods focus on improving performance under ideal conditions without natural or digital corruptions. The general absence of occlusions is assumed even for object-specific depth estimation. These methods are also vulnerable to adversarial attacks, which is a pertinent concern for their reliable deployment in robots and autonomous driving systems. We propose MIMDepth, a method that adapts masked image modeling (MIM) for self-supervised monocular depth estimation. While MIM has been used to learn generalizable features during pre-training, we show how it could be adapted for direct training of monocular depth estimation. Our experiments show that MIMDepth is more robust to noise, blur, weather conditions, digital artifacts, occlusions, as well as untargeted and targeted adversarial attacks.
comment: Accepted at 2023 IEEE International Conference on Robotics and Automation (ICRA)
♻ ☆ Matching entropy based disparity estimation from light field
A major challenge for matching-based depth estimation is to prevent mismatches in occlusion and smooth regions. An effective matching window satisfying three characteristics: texture richness, disparity consistency and anti-occlusion should be able to prevent mismatches to some extent. According to these characteristics, we propose matching entropy in the spatial domain of light field to measure the amount of correct information in a matching window, which provides the criterion for matching window selection. Based on matching entropy regularization, we establish an optimization model for depth estimation with a matching cost fidelity term. To find the optimum, we propose a two-step adaptive matching algorithm. First, the region type is adaptively determined to identify occluding, occluded, smooth and textured regions. Then, the matching entropy criterion is used to adaptively select the size and shape of matching windows, as well as the visible viewpoints. The two-step process can reduce mismatches and redundant calculations by selecting effective matching windows. The experimental results on synthetic and real data show that the proposed method can effectively improve the accuracy of depth estimation in occlusion and smooth regions and has strong robustness for different noise levels. Therefore, high-precision depth estimation from 4D light field data is achieved.
♻ ☆ Do Perceptually Aligned Gradients Imply Adversarial Robustness?
Adversarially robust classifiers possess a trait that non-robust models do not -- Perceptually Aligned Gradients (PAG). Their gradients with respect to the input align well with human perception. Several works have identified PAG as a byproduct of robust training, but none have considered it as a standalone phenomenon nor studied its own implications. In this work, we focus on this trait and test whether \emph{Perceptually Aligned Gradients imply Robustness}. To this end, we develop a novel objective to directly promote PAG in training classifiers and examine whether models with such gradients are more robust to adversarial attacks. Extensive experiments on multiple datasets and architectures validate that models with aligned gradients exhibit significant robustness, exposing the surprising bidirectional connection between PAG and robustness. Lastly, we show that better gradient alignment leads to increased robustness and harness this observation to boost the robustness of existing adversarial training techniques.
♻ ☆ PaCaNet: A Study on CycleGAN with Transfer Learning for Diversifying Fused Chinese Painting and Calligraphy
AI-Generated Content (AIGC) has recently gained a surge in popularity, powered by its high efficiency and consistency in production, and its capability of being customized and diversified. The cross-modality nature of the representation learning mechanism in most AIGC technology allows for more freedom and flexibility in exploring new types of art that would be impossible in the past. Inspired by the pictogram subset of Chinese characters, we proposed PaCaNet, a CycleGAN-based pipeline for producing novel artworks that fuse two different art types, traditional Chinese painting and calligraphy. In an effort to produce stable and diversified output, we adopted three main technical innovations: 1. Using one-shot learning to increase the creativity of pre-trained models and diversify the content of the fused images. 2. Controlling the preference over generated Chinese calligraphy by freezing randomly sampled parameters in pre-trained models. 3. Using a regularization method to encourage the models to produce images similar to Chinese paintings. Furthermore, we conducted a systematic study to explore the performance of PaCaNet in diversifying fused Chinese painting and calligraphy, which showed satisfying results. In conclusion, we provide a new direction of creating arts by fusing the visual information in paintings and the stroke features in Chinese calligraphy. Our approach creates a unique aesthetic experience rooted in the origination of Chinese hieroglyph characters. It is also a unique opportunity to delve deeper into traditional artwork and, in doing so, to create a meaningful impact on preserving and revitalizing traditional heritage.
♻ ☆ BIGRoC: Boosting Image Generation via a Robust Classifier
The interest of the machine learning community in image synthesis has grown significantly in recent years, with the introduction of a wide range of deep generative models and means for training them. In this work, we propose a general model-agnostic technique for improving the image quality and the distribution fidelity of generated images obtained by any generative model. Our method, termed BIGRoC (Boosting Image Generation via a Robust Classifier), is based on a post-processing procedure via the guidance of a given robust classifier and without a need for additional training of the generative model. Given a synthesized image, we propose to update it through projected gradient steps over the robust classifier to refine its recognition. We demonstrate this post-processing algorithm on various image synthesis methods and show a significant quantitative and qualitative improvement on CIFAR-10 and ImageNet. Surprisingly, although BIGRoC is the first model agnostic among refinement approaches and requires much less information, it outperforms competitive methods. Specifically, BIGRoC improves the image synthesis best performing diffusion model on ImageNet 128x128 by 14.81%, attaining an FID score of 2.53, and on 256x256 by 7.87%, achieving an FID of 3.63. Moreover, we conduct an opinion survey, according to which humans significantly prefer our method's outputs.
♻ ☆ Neural Shape Deformation Priors NeurIPS 2022
We present Neural Shape Deformation Priors, a novel method for shape manipulation that predicts mesh deformations of non-rigid objects from user-provided handle movements. State-of-the-art methods cast this problem as an optimization task, where the input source mesh is iteratively deformed to minimize an objective function according to hand-crafted regularizers such as ARAP. In this work, we learn the deformation behavior based on the underlying geometric properties of a shape, while leveraging a large-scale dataset containing a diverse set of non-rigid deformations. Specifically, given a source mesh and desired target locations of handles that describe the partial surface deformation, we predict a continuous deformation field that is defined in 3D space to describe the space deformation. To this end, we introduce transformer-based deformation networks that represent a shape deformation as a composition of local surface deformations. It learns a set of local latent codes anchored in 3D space, from which we can learn a set of continuous deformation functions for local surfaces. Our method can be applied to challenging deformations and generalizes well to unseen deformations. We validate our approach in experiments using the DeformingThing4D dataset, and compare to both classic optimization-based and recent neural network-based methods.
comment: NeurIPS 2022 Spotlight
♻ ☆ Open-Vocabulary Multi-Label Classification via Multi-Modal Knowledge Transfer AAAI 2023
Real-world recognition system often encounters the challenge of unseen labels. To identify such unseen labels, multi-label zero-shot learning (ML-ZSL) focuses on transferring knowledge by a pre-trained textual label embedding (e.g., GloVe). However, such methods only exploit single-modal knowledge from a language model, while ignoring the rich semantic information inherent in image-text pairs. Instead, recently developed open-vocabulary (OV) based methods succeed in exploiting such information of image-text pairs in object detection, and achieve impressive performance. Inspired by the success of OV-based methods, we propose a novel open-vocabulary framework, named multi-modal knowledge transfer (MKT), for multi-label classification. Specifically, our method exploits multi-modal knowledge of image-text pairs based on a vision and language pre-training (VLP) model. To facilitate transferring the image-text matching ability of VLP model, knowledge distillation is employed to guarantee the consistency of image and label embeddings, along with prompt tuning to further update the label embeddings. To further enable the recognition of multiple objects, a simple but effective two-stream module is developed to capture both local and global features. Extensive experimental results show that our method significantly outperforms state-of-the-art methods on public benchmark datasets. The source code is available at https://github.com/sunanhe/MKT.
comment: AAAI 2023 (Oral presentation paper). Updated version
♻ ☆ Diffusion-based Image Translation using Disentangled Style and Content Representation ICLR 2023
Diffusion-based image translation guided by semantic texts or a single target image has enabled flexible style transfer which is not limited to the specific domains. Unfortunately, due to the stochastic nature of diffusion models, it is often difficult to maintain the original content of the image during the reverse diffusion. To address this, here we present a novel diffusion-based unsupervised image translation method using disentangled style and content representation. Specifically, inspired by the splicing Vision Transformer, we extract intermediate keys of multihead self attention layer from ViT model and used them as the content preservation loss. Then, an image guided style transfer is performed by matching the [CLS] classification token from the denoised samples and target image, whereas additional CLIP loss is used for the text-driven style transfer. To further accelerate the semantic change during the reverse diffusion, we also propose a novel semantic divergence loss and resampling strategy. Our experimental results show that the proposed method outperforms state-of-the-art baseline models in both text-guided and image-guided translation tasks.
comment: ICLR 2023 camera ready
♻ ☆ Adversarial Driving: Attacking End-to-End Autonomous Driving
As research in deep neural networks advances, deep convolutional networks become promising for autonomous driving tasks. In particular, there is an emerging trend of employing end-to-end neural network models for autonomous driving. However, previous research has shown that deep neural network classifiers are vulnerable to adversarial attacks. While for regression tasks, the effect of adversarial attacks is not as well understood. In this research, we devise two white-box targeted attacks against end-to-end autonomous driving models. The driving system uses a regression model that takes an image as input and outputs the steering angle. Our attacks manipulate the behavior of the autonomous driving system by perturbing the input image. Both attacks can be initiated in real-time on CPUs without employing GPUs. The efficiency of the attacks is illustrated using experiments conducted in Udacity Simulator. Demo video: https://youtu.be/I0i8uN2oOP0.
comment: 7 pages, 5 figures
♻ ☆ Adversarial Detection: Attacking Object Detection in Real Time
Intelligent robots rely on object detection models to perceive the environment. Following advances in deep learning security it has been revealed that object detection models are vulnerable to adversarial attacks. However, prior research primarily focuses on attacking static images or offline videos. Therefore, it is still unclear if such attacks could jeopardize real-world robotic applications in dynamic environments. This paper bridges this gap by presenting the first real-time online attack against object detection models. We devise three attacks that fabricate bounding boxes for nonexistent objects at desired locations. The attacks achieve a success rate of about 90\% within about 20 iterations. The demo video is available at https://youtu.be/zJZ1aNlXsMU.
comment: 7 pages, 9 figures
♻ ☆ Combating Mode Collapse in GANs via Manifold Entropy Estimation AAAI'2023
Generative Adversarial Networks (GANs) have shown compelling results in various tasks and applications in recent years. However, mode collapse remains a critical problem in GANs. In this paper, we propose a novel training pipeline to address the mode collapse issue of GANs. Different from existing methods, we propose to generalize the discriminator as feature embedding and maximize the entropy of distributions in the embedding space learned by the discriminator. Specifically, two regularization terms, i.e., Deep Local Linear Embedding (DLLE) and Deep Isometric feature Mapping (DIsoMap), are designed to encourage the discriminator to learn the structural information embedded in the data, such that the embedding space learned by the discriminator can be well-formed. Based on the well-learned embedding space supported by the discriminator, a non-parametric entropy estimator is designed to efficiently maximize the entropy of embedding vectors, playing as an approximation of maximizing the entropy of the generated distribution. By improving the discriminator and maximizing the distance of the most similar samples in the embedding space, our pipeline effectively reduces the mode collapse without sacrificing the quality of generated samples. Extensive experimental results show the effectiveness of our method, which outperforms the GAN baseline, MaF-GAN on CelebA (9.13 vs. 12.43 in FID) and surpasses the recent state-of-the-art energy-based model on the ANIME-FACE dataset (2.80 vs. 2.26 in Inception score). The code is available at https://github.com/HaozheLiu-ST/MEE
comment: Accepted by AAAI'2023 (Oral); Code is released at https://github.com/HaozheLiu-ST/MEE
♻ ☆ What Makes Good Examples for Visual In-Context Learning?
Large-scale models trained on broad data have recently become the mainstream architecture in computer vision due to their strong generalization performance. In this paper, the main focus is on an emergent ability in large vision models, known as in-context learning, which allows inference on unseen tasks by conditioning on in-context examples (a.k.a.~prompt) without updating the model parameters. This concept has been well-known in natural language processing but has only been studied very recently for large vision models. We for the first time provide a comprehensive investigation on the impact of in-context examples in computer vision, and find that the performance is highly sensitive to the choice of in-context examples. To overcome the problem, we propose a prompt retrieval framework to automate the selection of in-context examples. Specifically, we present (1) an unsupervised prompt retrieval method based on nearest example search using an off-the-shelf model, and (2) a supervised prompt retrieval method, which trains a neural network to choose examples that directly maximize in-context learning performance. The results demonstrate that our methods can bring non-trivial improvements to visual in-context learning in comparison to the commonly-used random selection.
comment: code and models:https://github.com/ZhangYuanhan-AI/visual_prompt_retrieval
♻ ☆ GeoUDF: Surface Reconstruction from 3D Point Clouds via Geometry-guided Distance Representation
The recent neural implicit representation-based methods have greatly advanced the state of the art for solving the long-standing and challenging problem of reconstructing a discrete surface from a sparse point cloud. These methods generally learn either a binary occupancy or signed/unsigned distance field (SDF/UDF) as surface representation. However, existing SDF/UDF-based methods use neural networks to implicitly regress the distance in a purely data-driven manner, thus limiting the accuracy and generalizability to some extent. In contrast, we propose the first geometry-guided method for UDF and its gradient estimation that explicitly formulates the unsigned distance of a query point as the learnable affine averaging of its distances to the tangent planes of neighbouring points. Besides, we model the local geometric structure of the input point clouds by explicitly learning a quadratic polynomial for each point. This not only facilitates upsampling the input sparse point cloud but also naturally induces unoriented normal, which further augments UDF estimation. Finally, to extract triangle meshes from the predicted UDF we propose a customized edge-based marching cube module. We conduct extensive experiments and ablation studies to demonstrate the significant advantages of our method over state-of-the-art methods in terms of reconstruction accuracy, efficiency, and generalizability. The source code is publicly available at https://github.com/rsy6318/GeoUDF.
comment: 1 correct some unclear claim 2 add the results of DOG 3 redraw some figures
♻ ☆ Reversible Column Networks ICLR 2023
We propose a new neural network design paradigm Reversible Column Network (RevCol). The main body of RevCol is composed of multiple copies of subnetworks, named columns respectively, between which multi-level reversible connections are employed. Such architectural scheme attributes RevCol very different behavior from conventional networks: during forward propagation, features in RevCol are learned to be gradually disentangled when passing through each column, whose total information is maintained rather than compressed or discarded as other network does. Our experiments suggest that CNN-style RevCol models can achieve very competitive performances on multiple computer vision tasks such as image classification, object detection and semantic segmentation, especially with large parameter budget and large dataset. For example, after ImageNet-22K pre-training, RevCol-XL obtains 88.2% ImageNet-1K accuracy. Given more pre-training data, our largest model RevCol-H reaches 90.0% on ImageNet-1K, 63.8% APbox on COCO detection minival set, 61.0% mIoU on ADE20k segmentation. To our knowledge, it is the best COCO detection and ADE20k segmentation result among pure (static) CNN models. Moreover, as a general macro architecture fashion, RevCol can also be introduced into transformers or other neural networks, which is demonstrated to improve the performances in both computer vision and NLP tasks. We release code and models at https://github.com/megvii-research/RevCol
comment: Accepted by ICLR 2023
♻ ☆ Generative Category-Level Shape and Pose Estimation with Semantic Primitives
Empowering autonomous agents with 3D understanding for daily objects is a grand challenge in robotics applications. When exploring in an unknown environment, existing methods for object pose estimation are still not satisfactory due to the diversity of object shapes. In this paper, we propose a novel framework for category-level object shape and pose estimation from a single RGB-D image. To handle the intra-category variation, we adopt a semantic primitive representation that encodes diverse shapes into a unified latent space, which is the key to establish reliable correspondences between observed point clouds and estimated shapes. Then, by using a SIM(3)-invariant shape descriptor, we gracefully decouple the shape and pose of an object, thus supporting latent shape optimization of target objects in arbitrary poses. Extensive experiments show that the proposed method achieves SOTA pose estimation performance and better generalization in the real-world dataset. Code and video are available at https://zju3dv.github.io/gCasp.
comment: CoRL 2022, 18 pages, 13 figures, typos corrected, references added
♻ ☆ NeuS: Learning Neural Implicit Surfaces by Volume Rendering for Multi-view Reconstruction
We present a novel neural surface reconstruction method, called NeuS, for reconstructing objects and scenes with high fidelity from 2D image inputs. Existing neural surface reconstruction approaches, such as DVR and IDR, require foreground mask as supervision, easily get trapped in local minima, and therefore struggle with the reconstruction of objects with severe self-occlusion or thin structures. Meanwhile, recent neural methods for novel view synthesis, such as NeRF and its variants, use volume rendering to produce a neural scene representation with robustness of optimization, even for highly complex objects. However, extracting high-quality surfaces from this learned implicit representation is difficult because there are not sufficient surface constraints in the representation. In NeuS, we propose to represent a surface as the zero-level set of a signed distance function (SDF) and develop a new volume rendering method to train a neural SDF representation. We observe that the conventional volume rendering method causes inherent geometric errors (i.e. bias) for surface reconstruction, and therefore propose a new formulation that is free of bias in the first order of approximation, thus leading to more accurate surface reconstruction even without the mask supervision. Experiments on the DTU dataset and the BlendedMVS dataset show that NeuS outperforms the state-of-the-arts in high-quality surface reconstruction, especially for objects and scenes with complex structures and self-occlusion.
comment: 23 pages
♻ ☆ Pixel-Level Equalized Matching for Video Object Segmentation
Feature similarity matching, which transfers the information of the reference frame to the query frame, is a key component in semi-supervised video object segmentation. If surjective matching is adopted, background distractors can easily occur and degrade the performance. Bijective matching mechanisms try to prevent this by restricting the amount of information being transferred to the query frame, but have two limitations: 1) surjective matching cannot be fully leveraged as it is transformed to bijective matching at test time; and 2) test-time manual tuning is required for searching the optimal hyper-parameters. To overcome these limitations while ensuring reliable information transfer, we introduce an equalized matching mechanism. To prevent the reference frame information from being overly referenced, the potential contribution to the query frame is equalized by simply applying a softmax operation along with the query. On public benchmark datasets, our proposed approach achieves a comparable performance to state-of-the-art methods.
♻ ☆ What's in a Decade? Transforming Faces Through Time
How can one visually characterize people in a decade? In this work, we assemble the Faces Through Time dataset, which contains over a thousand portrait images from each decade, spanning the 1880s to the present day. Using our new dataset, we present a framework for resynthesizing portrait images across time, imagining how a portrait taken during a particular decade might have looked like, had it been taken in other decades. Our framework optimizes a family of per-decade generators that reveal subtle changes that differentiate decade--such as different hairstyles or makeup--while maintaining the identity of the input portrait. Experiments show that our method is more effective in resynthesizing portraits across time compared to state-of-the-art image-to-image translation methods, as well as attribute-based and language-guided portrait editing models. Our code and data will be available at https://facesthroughtime.github.io
comment: Project Page: https://facesthroughtime.github.io
♻ ☆ DisDiff: Unsupervised Disentanglement of Diffusion Probabilistic Models
In this paper, targeting to understand the underlying explainable factors behind observations and modeling the conditional generation process on these factors, we propose a new task, disentanglement of diffusion probabilistic models (DPMs), to take advantage of the remarkable modeling ability of DPMs. To tackle this task, we further devise an unsupervised approach named DisDiff. For the first time, we achieve disentangled representation learning in the framework of diffusion probabilistic models. Given a pre-trained DPM, DisDiff can automatically discover the inherent factors behind the image data and disentangle the gradient fields of DPM into sub-gradient fields, each conditioned on the representation of each discovered factor. We propose a novel Disentangling Loss for DisDiff to facilitate the disentanglement of the representation and sub-gradients. The extensive experiments on synthetic and real-world datasets demonstrate the effectiveness of DisDiff.
♻ ☆ Expanding the Deployment Envelope of Behavior Prediction via Adaptive Meta-Learning ICRA 2023
Learning-based behavior prediction methods are increasingly being deployed in real-world autonomous systems, e.g., in fleets of self-driving vehicles, which are beginning to commercially operate in major cities across the world. Despite their advancements, however, the vast majority of prediction systems are specialized to a set of well-explored geographic regions or operational design domains, complicating deployment to additional cities, countries, or continents. Towards this end, we present a novel method for efficiently adapting behavior prediction models to new environments. Our approach leverages recent advances in meta-learning, specifically Bayesian regression, to augment existing behavior prediction models with an adaptive layer that enables efficient domain transfer via offline fine-tuning, online adaptation, or both. Experiments across multiple real-world datasets demonstrate that our method can efficiently adapt to a variety of unseen environments.
comment: 12 pages, 13 figures, 2 tables. To appear at ICRA 2023
♻ ☆ Quality Not Quantity: On the Interaction between Dataset Design and Robustness of CLIP NeurIPS 2022
Web-crawled datasets have enabled remarkable generalization capabilities in recent image-text models such as CLIP (Contrastive Language-Image pre-training) or Flamingo, but little is known about the dataset creation processes. In this work, we introduce a testbed of six publicly available data sources - YFCC, LAION, Conceptual Captions, WIT, RedCaps, Shutterstock - to investigate how pre-training distributions induce robustness in CLIP. We find that the performance of the pre-training data varies substantially across distribution shifts, with no single data source dominating. Moreover, we systematically study the interactions between these data sources and find that combining multiple sources does not necessarily yield better models, but rather dilutes the robustness of the best individual data source. We complement our empirical findings with theoretical insights from a simple setting, where combining the training data also results in diluted robustness. In addition, our theoretical model provides a candidate explanation for the success of the CLIP-based data filtering technique recently employed in the LAION dataset. Overall our results demonstrate that simply gathering a large amount of data from the web is not the most effective way to build a pre-training dataset for robust generalization, necessitating further study into dataset design. Code is available at https://github.com/mlfoundations/clip_quality_not_quantity.
comment: Oral paper at NeurIPS 2022
♻ ☆ PV3D: A 3D Generative Model for Portrait Video Generation ICLR2023
Recent advances in generative adversarial networks (GANs) have demonstrated the capabilities of generating stunning photo-realistic portrait images. While some prior works have applied such image GANs to unconditional 2D portrait video generation and static 3D portrait synthesis, there are few works successfully extending GANs for generating 3D-aware portrait videos. In this work, we propose PV3D, the first generative framework that can synthesize multi-view consistent portrait videos. Specifically, our method extends the recent static 3D-aware image GAN to the video domain by generalizing the 3D implicit neural representation to model the spatio-temporal space. To introduce motion dynamics to the generation process, we develop a motion generator by stacking multiple motion layers to generate motion features via modulated convolution. To alleviate motion ambiguities caused by camera/human motions, we propose a simple yet effective camera condition strategy for PV3D, enabling both temporal and multi-view consistent video generation. Moreover, PV3D introduces two discriminators for regularizing the spatial and temporal domains to ensure the plausibility of the generated portrait videos. These elaborated designs enable PV3D to generate 3D-aware motion-plausible portrait videos with high-quality appearance and geometry, significantly outperforming prior works. As a result, PV3D is able to support many downstream applications such as animating static portraits and view-consistent video motion editing. Code and models are released at https://showlab.github.io/pv3d.
comment: Accepted to ICLR2023, Project Page https://showlab.github.io/pv3d
♻ ☆ GaitSADA: Self-Aligned Domain Adaptation for mmWave Gait Recognition
mmWave radar-based gait recognition is a novel user identification method that captures human gait biometrics from mmWave radar return signals. This technology offers privacy protection and is resilient to weather and lighting conditions. However, its generalization performance is yet unknown and limits its practical deployment. To address this problem, in this paper, a non-synthetic dataset is collected and analyzed to reveal the presence of spatial and temporal domain shifts in mmWave gait biometric data, which significantly impacts identification accuracy. To address this issue, a novel self-aligned domain adaptation method called GaitSADA is proposed. GaitSADA improves system generalization performance by using a two-stage semi-supervised model training approach. The first stage uses semi-supervised contrastive learning and the second stage uses semi-supervised consistency training with centroid alignment. Extensive experiments show that GaitSADA outperforms representative domain adaptation methods by an average of 15.41% in low data regimes.
comment: Submitted to ACM Transactions on Sensor Networks (TOSN)
♻ ☆ Automated Cardiac Resting Phase Detection Targeted on the Right Coronary Artery
Static cardiac imaging such as late gadolinium enhancement, mapping, or 3-D coronary angiography require prior information, e.g., the phase during a cardiac cycle with least motion, called resting phase (RP). The purpose of this work is to propose a fully automated framework that allows the detection of the right coronary artery (RCA) RP within CINE series. The proposed prototype system consists of three main steps. First, the localization of the regions of interest (ROI) is performed. Second, the cropped ROI series are taken for tracking motions over all time points. Third, the output motion values are used to classify RPs. In this work, we focused on the detection of the area with the outer edge of the cross-section of the RCA as our target. The proposed framework was evaluated on 102 clinically acquired dataset at 1.5T and 3T. The automatically classified RPs were compared with the reference RPs annotated manually by a expert for testing the robustness and feasibility of the framework. The predicted RCA RPs showed high agreement with the experts annotated RPs with 92.7% accuracy, 90.5% sensitivity and 95.0% specificity for the unseen study dataset. The mean absolute difference of the start and end RP was 13.6 $\pm$ 18.6 ms for the validation study dataset (n=102). In this work, automated RP detection has been introduced by the proposed framework and demonstrated feasibility, robustness, and applicability for static imaging acquisitions.
comment: Accepted for publication at the Journal of Machine Learning for Biomedical Imaging (MELBA) https://melba-journal.org/2023:001
♻ ☆ E-VFIA : Event-Based Video Frame Interpolation with Attention ICRA 2023
Video frame interpolation (VFI) is a fundamental vision task that aims to synthesize several frames between two consecutive original video images. Most algorithms aim to accomplish VFI by using only keyframes, which is an ill-posed problem since the keyframes usually do not yield any accurate precision about the trajectories of the objects in the scene. On the other hand, event-based cameras provide more precise information between the keyframes of a video. Some recent state-of-the-art event-based methods approach this problem by utilizing event data for better optical flow estimation to interpolate for video frame by warping. Nonetheless, those methods heavily suffer from the ghosting effect. On the other hand, some of kernel-based VFI methods that only use frames as input, have shown that deformable convolutions, when backed up with transformers, can be a reliable way of dealing with long-range dependencies. We propose event-based video frame interpolation with attention (E-VFIA), as a lightweight kernel-based method. E-VFIA fuses event information with standard video frames by deformable convolutions to generate high quality interpolated frames. The proposed method represents events with high temporal resolution and uses a multi-head self-attention mechanism to better encode event-based information, while being less vulnerable to blurring and ghosting artifacts; thus, generating crispier frames. The simulation results show that the proposed technique outperforms current state-of-the-art methods (both frame and event-based) with a significantly smaller model size.
comment: Accepted to 2023 IEEE International Conference on Robotics and Automation (ICRA 2023)
♻ ☆ Interaction Transformer for Human Reaction Generation
We address the challenging task of human reaction generation, which aims to generate a corresponding reaction based on an input action. Most of the existing works do not focus on generating and predicting the reaction and cannot generate the motion when only the action is given as input. To address this limitation, we propose a novel interaction Transformer (InterFormer) consisting of a Transformer network with both temporal and spatial attention. Specifically, temporal attention captures the temporal dependencies of the motion of both characters and of their interaction, while spatial attention learns the dependencies between the different body parts of each character and those which are part of the interaction. Moreover, we propose using graphs to increase the performance of spatial attention via an interaction distance module that helps focus on nearby joints from both characters. Extensive experiments on the SBU interaction, K3HI, and DuetDance datasets demonstrate the effectiveness of InterFormer. Our method is general and can be used to generate more complex and long-term interactions. We also provide videos of generated reactions and the code with pre-trained models at https://github.com/CRISTAL-3DSAM/InterFormer
♻ ☆ DDM-NET: End-to-end learning of keypoint feature Detection, Description and Matching for 3D localization
In this paper, we propose an end-to-end framework that jointly learns keypoint detection, descriptor representation and cross-frame matching for the task of image-based 3D localization. Prior art has tackled each of these components individually, purportedly aiming to alleviate difficulties in effectively train a holistic network. We design a self-supervised image warping correspondence loss for both feature detection and matching, a weakly-supervised epipolar constraints loss on relative camera pose learning, and a directional matching scheme that detects key-point features in a source image and performs coarse-to-fine correspondence search on the target image. We leverage this framework to enforce cycle consistency in our matching module. In addition, we propose a new loss to robustly handle both definite inlier/outlier matches and less-certain matches. The integration of these learning mechanisms enables end-to-end training of a single network performing all three localization components. Bench-marking our approach on public data-sets, exemplifies how such an end-to-end framework is able to yield more accurate localization that out-performs both traditional methods as well as state-of-the-art weakly supervised methods.
Information Retrieval 8
☆ 'Generative CI' through Collective Response Systems
How can many people (who may disagree) come together to answer a question or make a decision? "Collective response systems" are a type of generative collective intelligence (CI) facilitation process meant to address this challenge. They enable a form of "generative voting", where both the votes, and the choices of what to vote on, are provided by the group. Such systems overcome the traditional limitations of polling, town halls, standard voting, referendums, etc. The generative CI outputs of collective response systems can also be chained together into iterative "collective dialogues", analogously to some kinds of generative AI. Technical advances across domains including recommender systems, language models, and human-computer interaction have led to the development of innovative and scalable collective response systems. For example, Polis has been used around the world to support policy-making at different levels of government, and Remesh has been used by the UN to understand the challenges and needs of ordinary people across war-torn countries. This paper aims to develop a shared language by defining the structure, processes, properties, and principles of such systems. Collective response systems allow non-confrontational exploration of divisive issues, help identify common ground, and elicit insights from those closest to the issues. As a result, they can help overcome gridlock around conflict and governance challenges, increase trust, and develop mandates. Continued progress toward their development and adoption could help revitalize democracies, reimagine corporate governance, transform conflict, and govern powerful AI systems -- both as a complement to deeper deliberative democratic processes and as an option where deeper processes are not applicable or possible.
comment: 5 pages
☆ Faster maximal clique enumeration in large real-world link streams
Link streams offer a good model for representing interactions over time. They consist of links $(b,e,u,v)$, where $u$ and $v$ are vertices interacting during the whole time interval $[b,e]$. In this paper, we deal with the problem of enumerating maximal cliques in link streams. A clique is a pair $(C,[t_0,t_1])$, where $C$ is a set of vertices that all interact pairwise during the full interval $[t_0,t_1]$. It is maximal when neither its set of vertices nor its time interval can be increased. Some of the main works solving this problem are based on the famous Bron-Kerbosch algorithm for enumerating maximal cliques in graphs. We take this idea as a starting point to propose a new algorithm which matches the cliques of the instantaneous graphs formed by links existing at a given time $t$ to the maximal cliques of the link stream. We prove its validity and compute its complexity, which is better than the state-of-the art ones in many cases of interest. We also study the output-sensitive complexity, which is close to the output size, thereby showing that our algorithm is efficient. To confirm this, we perform experiments on link streams used in the state of the art, and on massive link streams, up to 100 million links. In all cases our algorithm is faster, mostly by a factor of at least 10 and up to a factor of $10^4$. Moreover, it scales to massive link streams for which the existing algorithms are not able to provide the solution.
comment: 27 pages, 6 figure, 5 tables, submitted to Journal of Graph Algorithms and Applications
☆ Fairness-aware Cross-Domain Recommendation
Cross-Domain Recommendation (CDR) is an effective way to alleviate the cold-start problem. However, previous work severely ignores fairness and bias when learning the mapping function, which is used to obtain the representations for fresh users in the target domain. To study this problem, in this paper, we propose a Fairness-aware Cross-Domain Recommendation model, called FairCDR. Our method achieves user-oriented group fairness by learning the fairness-aware mapping function. Since the overlapping data are quite limited and distributionally biased, FairCDR leverages abundant non-overlapping users and interactions to help alleviate these problems. Considering that each individual has different influence on model fairness, we propose a new reweighing method based on Influence Function (IF) to reduce unfairness while maintaining recommendation accuracy. Extensive experiments are conducted to demonstrate the effectiveness of our model.
☆ Unsupervised Entity Alignment for Temporal Knowledge Graphs WWW
Entity alignment (EA) is a fundamental data integration task that identifies equivalent entities between different knowledge graphs (KGs). Temporal Knowledge graphs (TKGs) extend traditional knowledge graphs by introducing timestamps, which have received increasing attention. State-of-the-art time-aware EA studies have suggested that the temporal information of TKGs facilitates the performance of EA. However, existing studies have not thoroughly exploited the advantages of temporal information in TKGs. Also, they perform EA by pre-aligning entity pairs, which can be labor-intensive and thus inefficient. In this paper, we present DualMatch which effectively fuses the relational and temporal information for EA. DualMatch transfers EA on TKGs into a weighted graph matching problem. More specifically, DualMatch is equipped with an unsupervised method, which achieves EA without necessitating seed alignment. DualMatch has two steps: (i) encoding temporal and relational information into embeddings separately using a novel label-free encoder, Dual-Encoder; and (ii) fusing both information and transforming it into alignment using a novel graph-matching-based decoder, GM-Decoder. DualMatch is able to perform EA on TKGs with or without supervision, due to its capability of effectively capturing temporal information. Extensive experiments on three real-world TKG datasets offer the insight that DualMatch outperforms the state-of-the-art methods in terms of H@1 by 2.4% - 10.7% and MRR by 1.7% - 7.6%, respectively.
comment: Accepted by The Web Conference (WWW) 2023 Research Track
♻ ☆ Towards Disentangling Relevance and Bias in Unbiased Learning to Rank
Unbiased learning to rank (ULTR) studies the problem of mitigating various biases from implicit user feedback data such as clicks, and has been receiving considerable attention recently. A popular ULTR approach for real-world applications uses a two-tower architecture, where click modeling is factorized into a relevance tower with regular input features, and a bias tower with bias-relevant inputs such as the position of a document. A successful factorization will allow the relevance tower to be exempt from biases. In this work, we identify a critical issue that existing ULTR methods ignored - the bias tower can be confounded with the relevance tower via the underlying true relevance. In particular, the positions were determined by the logging policy, i.e., the previous production model, which would possess relevance information. We give both theoretical analysis and empirical results to show the negative effects on relevance tower due to such a correlation. We then propose three methods to mitigate the negative confounding effects by better disentangling relevance and bias. Empirical results on both controlled public datasets and a large-scale industry dataset show the effectiveness of the proposed approaches.
♻ ☆ CitationSum: Citation-aware Graph Contrastive Learning for Scientific Paper Summarization WWW2023
The citation graph is essential for generating high-quality summaries of scientific papers, in which references of a scientific paper and their correlations provide extra knowledge for understanding its background and main contributions. Despite the promising role of the citation graph, effectively incorporating it still remains a big challenge, given the difficulty of accurately identifying and leveraging relevant contents in references for a source paper, as well as modelling their correlations of different intensities. Existing methods either ignore or utilize only abstracts indiscriminately from references, failing to tackle the challenge mentioned above. To fill the gap, we propose a novel citation-aware scientific paper summarization framework based on the citation graph, with the ability to accurately locate and incorporate the salient contents from references, as well as capture varying relevance between source papers and their references. Specifically, we first build a domain-specific dataset PubMedCite with about 192K biomedical scientific papers and a large citation graph preserving 917K citation relationships between them. It is characterized by preserving the salient contents extracted from full texts of references, and the weighted correlation between the salient contents of references and the source paper. Based on it, we design a self-supervised citation-aware summarization framework (CitationSum) with graph contrastive learning, which boosts the summarization generation by efficiently fusing the salient information in references with source paper contents under the guidance of their correlations. Experimental results show that our model outperforms the state-of-the-art methods, due to efficiently leveraging the information of references and citation correlations.
comment: accepted to WWW2023
♻ ☆ Pessimistic Off-Policy Optimization for Learning to Rank
Off-policy learning is a framework for optimizing policies without deploying them, using data collected by another policy. In recommender systems, this is especially challenging due to the imbalance in logged data: some items are recommended and thus logged more frequently than others. This is further perpetuated when recommending a list of items, as the action space is combinatorial. To address this challenge, we study pessimistic off-policy optimization for learning to rank. The key idea is to compute lower confidence bounds on parameters of click models and then return the list with the highest pessimistic estimate of its value. This approach is computationally efficient and we analyze it. We study its Bayesian and frequentist variants, and overcome the limitation of unknown prior by incorporating empirical Bayes. To show the empirical effectiveness of our approach, we compare it to off-policy optimizers that use inverse propensity scores or neglect uncertainty. Our approach outperforms all baselines, is robust, and is also general.
♻ ☆ Human Mobility Prediction with Causal and Spatial-constrained Multi-task Network
Modeling human mobility helps to understand how people are accessing resources and physically contacting with each other in cities, and thus contributes to various applications such as urban planning, epidemic control, and location-based advertisement. Next location prediction is one decisive task in individual human mobility modeling and is usually viewed as sequence modeling, solved with Markov or RNN-based methods. However, the existing models paid little attention to the logic of individual travel decisions and the reproducibility of the collective behavior of population. To this end, we propose a Causal and Spatial-constrained Long and Short-term Learner (CSLSL) for next location prediction. CSLSL utilizes a causal structure based on multi-task learning to explicitly model the "\textit{when$\rightarrow$what$\rightarrow$where}", a.k.a. "\textit{time$\rightarrow$activity$\rightarrow$location}" decision logic. We next propose a spatial-constrained loss function as an auxiliary task, to ensure the consistency between the predicted and actual spatial distribution of travelers' destinations. Moreover, CSLSL adopts modules named Long and Short-term Capturer (LSC) to learn the transition regularities across different time spans. Extensive experiments on three real-world datasets show promising performance improvements of CSLSL over baselines and confirm the effectiveness of introducing the causality and consistency constraints. The implementation is available at https://github.com/urbanmobility/CSLSL.
comment: Updated version
Machine Learning 144
☆ Improving Few-Shot Generalization by Exploring and Exploiting Auxiliary Data
Few-shot learning involves learning an effective model from only a few labeled datapoints. The use of a small training set makes it difficult to avoid overfitting but also makes few-shot learning applicable to many important real-world settings. In this work, we focus on Few-shot Learning with Auxiliary Data (FLAD), a training paradigm that assumes access to auxiliary data during few-shot learning in hopes of improving generalization. Introducing auxiliary data during few-shot learning leads to essential design choices where hand-designed heuristics can lead to sub-optimal performance. In this work, we focus on automated sampling strategies for FLAD and relate them to the explore-exploit dilemma that is central in multi-armed bandit settings. Based on this connection we propose two algorithms -- EXP3-FLAD and UCB1-FLAD -- and compare them with methods that either explore or exploit, finding that the combination of exploration and exploitation is crucial. Using our proposed algorithms to train T5 yields a 9% absolute improvement over the explicitly multi-task pre-trained T0 model across 11 datasets.
comment: 19 pages, 7 figures, code available at https://github.com/alon-albalak/FLAD
☆ Efficient Multi-Task Reinforcement Learning via Selective Behavior Sharing
The ability to leverage shared behaviors between tasks is critical for sample-efficient multi-task reinforcement learning (MTRL). While prior methods have primarily explored parameter and data sharing, direct behavior-sharing has been limited to task families requiring similar behaviors. Our goal is to extend the efficacy of behavior-sharing to more general task families that could require a mix of shareable and conflicting behaviors. Our key insight is an agent's behavior across tasks can be used for mutually beneficial exploration. To this end, we propose a simple MTRL framework for identifying shareable behaviors over tasks and incorporating them to guide exploration. We empirically demonstrate how behavior sharing improves sample efficiency and final performance on manipulation and navigation MTRL tasks and is even complementary to parameter sharing. Result videos are available at https://sites.google.com/view/qmp-mtrl.
☆ Stable Target Field for Reduced Variance Score Estimation in Diffusion Models ICLR 2023
Diffusion models generate samples by reversing a fixed forward diffusion process. Despite already providing impressive empirical results, these diffusion models algorithms can be further improved by reducing the variance of the training targets in their denoising score-matching objective. We argue that the source of such variance lies in the handling of intermediate noise-variance scales, where multiple modes in the data affect the direction of reverse paths. We propose to remedy the problem by incorporating a reference batch which we use to calculate weighted conditional scores as more stable training targets. We show that the procedure indeed helps in the challenging intermediate regime by reducing (the trace of) the covariance of training targets. The new stable targets can be seen as trading bias for reduced variance, where the bias vanishes with increasing reference batch size. Empirically, we show that the new objective improves the image quality, stability, and training speed of various popular diffusion models across datasets with both general ODE and SDE solvers. When used in combination with EDM, our method yields a current SOTA FID of 1.90 with 35 network evaluations on the unconditional CIFAR-10 generation task. The code is available at https://github.com/Newbeeer/stf
comment: Accepted by ICLR 2023. Code available at: https://github.com/Newbeeer/stf
☆ Robust Fitted-Q-Evaluation and Iteration under Sequentially Exogenous Unobserved Confounders
Offline reinforcement learning is important in domains such as medicine, economics, and e-commerce where online experimentation is costly, dangerous or unethical, and where the true model is unknown. However, most methods assume all covariates used in the behavior policy's action decisions are observed. This untestable assumption may be incorrect. We study robust policy evaluation and policy optimization in the presence of unobserved confounders. We assume the extent of possible unobserved confounding can be bounded by a sensitivity model, and that the unobserved confounders are sequentially exogenous. We propose and analyze an (orthogonalized) robust fitted-Q-iteration that uses closed-form solutions of the robust Bellman operator to derive a loss minimization problem for the robust Q function. Our algorithm enjoys the computational ease of fitted-Q-iteration and statistical improvements (reduced dependence on quantile estimation error) from orthogonalization. We provide sample complexity bounds, insights, and show effectiveness in simulations.
☆ Graph Neural Operators for Classification of Spatial Transcriptomics Data
The inception of spatial transcriptomics has allowed improved comprehension of tissue architectures and the disentanglement of complex underlying biological, physiological, and pathological processes through their positional contexts. Recently, these contexts, and by extension the field, have seen much promise and elucidation with the application of graph learning approaches. In particular, neural operators have risen in regards to learning the mapping between infinite-dimensional function spaces. With basic to deep neural network architectures being data-driven, i.e. dependent on quality data for prediction, neural operators provide robustness by offering generalization among different resolutions despite low quality data. Graph neural operators are a variant that utilize graph networks to learn this mapping between function spaces. The aim of this research is to identify robust machine learning architectures that integrate spatial information to predict tissue types. Under this notion, we propose a study incorporating various graph neural network approaches to validate the efficacy of applying neural operators towards prediction of brain regions in mouse brain tissue samples as a proof of concept towards our purpose. We were able to achieve an F1 score of nearly 72% for the graph neural operator approach which outperformed all baseline and other graph network approaches.
comment: 5 pages, 8 figures, 1 table. To appear in conference proceedings and journal publications
☆ Epic-Sounds: A Large-scale Dataset of Actions That Sound
We introduce EPIC-SOUNDS, a large-scale dataset of audio annotations capturing temporal extents and class labels within the audio stream of the egocentric videos. We propose an annotation pipeline where annotators temporally label distinguishable audio segments and describe the action that could have caused this sound. We identify actions that can be discriminated purely from audio, through grouping these free-form descriptions of audio into classes. For actions that involve objects colliding, we collect human annotations of the materials of these objects (e.g. a glass object being placed on a wooden surface), which we verify from visual labels, discarding ambiguities. Overall, EPIC-SOUNDS includes 78.4k categorised segments of audible events and actions, distributed across 44 classes as well as 39.2k non-categorised segments. We train and evaluate two state-of-the-art audio recognition models on our dataset, highlighting the importance of audio-only labels and the limitations of current models to recognise actions that sound.
comment: 6 pages, 4 figures
☆ Deep Dependency Networks for Multi-Label Classification
We propose a simple approach which combines the strengths of probabilistic graphical models and deep learning architectures for solving the multi-label classification task, focusing specifically on image and video data. First, we show that the performance of previous approaches that combine Markov Random Fields with neural networks can be modestly improved by leveraging more powerful methods such as iterative join graph propagation, integer linear programming, and $\ell_1$ regularization-based structure learning. Then we propose a new modeling framework called deep dependency networks, which augments a dependency network, a model that is easy to train and learns more accurate dependencies but is limited to Gibbs sampling for inference, to the output layer of a neural network. We show that despite its simplicity, jointly learning this new architecture yields significant improvements in performance over the baseline neural network. In particular, our experimental evaluation on three video activity classification datasets: Charades, Textually Annotated Cooking Scenes (TACoS), and Wetlab, and three multi-label image classification datasets: MS-COCO, PASCAL VOC, and NUS-WIDE show that deep dependency networks are almost always superior to pure neural architectures that do not use dependency networks.
☆ Training Normalizing Flows with the Precision-Recall Divergence
Generative models can have distinct mode of failures like mode dropping and low quality samples, which cannot be captured by a single scalar metric. To address this, recent works propose evaluating generative models using precision and recall, where precision measures quality of samples and recall measures the coverage of the target distribution. Although a variety of discrepancy measures between the target and estimated distribution are used to train generative models, it is unclear what precision-recall trade-offs are achieved by various choices of the discrepancy measures. In this paper, we show that achieving a specified precision-recall trade-off corresponds to minimising -divergences from a family we call the {\em PR-divergences }. Conversely, any -divergence can be written as a linear combination of PR-divergences and therefore correspond to minimising a weighted precision-recall trade-off. Further, we propose a novel generative model that is able to train a normalizing flow to minimise any -divergence, and in particular, achieve a given precision-recall trade-off.
☆ Efficient Meta-Learning via Error-based Context Pruning for Implicit Neural Representations
We introduce an efficient optimization-based meta-learning technique for learning large-scale implicit neural representations (INRs). Our main idea is designing an online selection of context points, which can significantly reduce memory requirements for meta-learning in any established setting. By doing so, we expect additional memory savings which allows longer per-signal adaptation horizons (at a given memory budget), leading to better meta-initializations by reducing myopia and, more crucially, enabling learning on high-dimensional signals. To implement such context pruning, our technical novelty is three-fold. First, we propose a selection scheme that adaptively chooses a subset at each adaptation step based on the predictive error, leading to the modeling of the global structure of the signal in early steps and enabling the later steps to capture its high-frequency details. Second, we counteract any possible information loss from context pruning by minimizing the parameter distance to a bootstrapped target model trained on a full context set. Finally, we suggest using the full context set with a gradient scaling scheme at test-time. Our technique is model-agnostic, intuitive, and straightforward to implement, showing significant reconstruction improvements for a wide range of signals. Code is available at https://github.com/jihoontack/ECoP
comment: 19 pages
GFlowNets for AI-Driven Scientific Discovery
Tackling the most pressing problems for humanity, such as the climate crisis and the threat of global pandemics, requires accelerating the pace of scientific discovery. While science has traditionally relied on trial and error and even serendipity to a large extent, the last few decades have seen a surge of data-driven scientific discoveries. However, in order to truly leverage large-scale data sets and high-throughput experimental setups, machine learning methods will need to be further improved and better integrated in the scientific discovery pipeline. A key challenge for current machine learning methods in this context is the efficient exploration of very large search spaces, which requires techniques for estimating reducible (epistemic) uncertainty and generating sets of diverse and informative experiments to perform. This motivated a new probabilistic machine learning framework called GFlowNets, which can be applied in the modeling, hypotheses generation and experimental design stages of the experimental science loop. GFlowNets learn to sample from a distribution given indirectly by a reward function corresponding to an unnormalized probability, which enables sampling diverse, high-reward candidates. GFlowNets can also be used to form efficient and amortized Bayesian posterior estimators for causal models conditioned on the already acquired experimental data. Having such posterior models can then provide estimators of epistemic uncertainty and information gain that can drive an experimental design policy. Altogether, here we will argue that GFlowNets can become a valuable tool for AI-driven scientific discovery, especially in scenarios of very large candidate spaces where we have access to cheap but inaccurate measurements or to expensive but accurate measurements. This is a common setting in the context of drug and material discovery, which we use as examples throughout the paper.
comment: 26 pages, 5 figures
☆ Uniswap Liquidity Provision: An Online Learning Approach
Decentralized Exchanges (DEXs) are new types of marketplaces leveraging Blockchain technology. They allow users to trade assets with Automatic Market Makers (AMM), using funds provided by liquidity providers, removing the need for order books. One such DEX, Uniswap v3, allows liquidity providers to allocate funds more efficiently by specifying an active price interval for their funds. This introduces the problem of finding an optimal strategy for choosing price intervals. We formalize this problem as an online learning problem with non-stochastic rewards. We use regret-minimization methods to show a liquidity provision strategy that guarantees a lower bound on the reward. This is true even for non-stochastic changes to asset pricing, and we express this bound in terms of the trading volume.
☆ Two for One: Diffusion Models and Force Fields for Coarse-Grained Molecular Dynamics
Coarse-grained (CG) molecular dynamics enables the study of biological processes at temporal and spatial scales that would be intractable at an atomistic resolution. However, accurately learning a CG force field remains a challenge. In this work, we leverage connections between score-based generative models, force fields and molecular dynamics to learn a CG force field without requiring any force inputs during training. Specifically, we train a diffusion generative model on protein structures from molecular dynamics simulations, and we show that its score function approximates a force field that can directly be used to simulate CG molecular dynamics. While having a vastly simplified training setup compared to previous work, we demonstrate that our approach leads to improved performance across several small- to medium-sized protein simulations, reproducing the CG equilibrium distribution, and preserving dynamics of all-atom simulations such as protein folding events.
☆ MB-DECTNet: A Model-Based Unrolled Network for Accurate 3D DECT Reconstruction
Numerous dual-energy CT (DECT) techniques have been developed in the past few decades. Dual-energy CT (DECT) statistical iterative reconstruction (SIR) has demonstrated its potential for reducing noise and increasing accuracy. Our lab proposed a joint statistical DECT algorithm for stopping power estimation and showed that it outperforms competing image-based material-decomposition methods. However, due to its slow convergence and the high computational cost of projections, the elapsed time of 3D DECT SIR is often not clinically acceptable. Therefore, to improve its convergence, we have embedded DECT SIR into a deep learning model-based unrolled network for 3D DECT reconstruction (MB-DECTNet) that can be trained in an end-to-end fashion. This deep learning-based method is trained to learn the shortcuts between the initial conditions and the stationary points of iterative algorithms while preserving the unbiased estimation property of model-based algorithms. MB-DECTNet is formed by stacking multiple update blocks, each of which consists of a data consistency layer (DC) and a spatial mixer layer, where the spatial mixer layer is the shrunken U-Net, and the DC layer is a one-step update of an arbitrary traditional iterative method. Although the proposed network can be combined with numerous iterative DECT algorithms, we demonstrate its performance with the dual-energy alternating minimization (DEAM). The qualitative result shows that MB-DECTNet with DEAM significantly reduces noise while increasing the resolution of the test image. The quantitative result shows that MB-DECTNet has the potential to estimate attenuation coefficients accurately as traditional statistical algorithms but with a much lower computational cost.
☆ Machine Learning for Visualization Recommendation Systems: Open Challenges and Future Directions
Visualization Recommendation Systems (VRS) are a novel and challenging field of study, whose aim is to automatically generate insightful visualizations from data, to support non-expert users in the process of information discovery. Despite its enormous application potential in the era of big data, progress in this area of research is being held back by several obstacles among which are the absence of standardized datasets to train recommendation algorithms, and the difficulty in defining quantitative criteria to assess the effectiveness of the generated plots. In this paper, we aim not only to summarize the state-of-the-art of VRS, but also to outline promising future research directions.
comment: 9 pages, 2 figures
☆ Automatically Marginalized MCMC in Probabilistic Programming
Hamiltonian Monte Carlo (HMC) is a powerful algorithm to sample latent variables from Bayesian models. The advent of probabilistic programming languages (PPLs) frees users from writing inference algorithms and lets users focus on modeling. However, many models are difficult for HMC to solve directly, which often require tricks like model reparameterization. We are motivated by the fact that many of those models could be simplified by marginalization. We propose to use automatic marginalization as part of the sampling process using HMC in a graphical model extracted from a PPL, which substantially improves sampling from real-world hierarchical models.
☆ Graph Neural Network Based Surrogate Model of Physics Simulations for Geometry Design
Computational Intelligence (CI) techniques have shown great potential as a surrogate model of expensive physics simulation, with demonstrated ability to make fast predictions, albeit at the expense of accuracy in some cases. For many scientific and engineering problems involving geometrical design, it is desirable for the surrogate models to precisely describe the change in geometry and predict the consequences. In that context, we develop graph neural networks (GNNs) as fast surrogate models for physics simulation, which allow us to directly train the models on 2/3D geometry designs that are represented by an unstructured mesh or point cloud, without the need for any explicit or hand-crafted parameterization. We utilize an encoder-processor-decoder-type architecture which can flexibly make prediction at both node level and graph level. The performance of our proposed GNN-based surrogate model is demonstrated on 2 example applications: feature designs in the domain of additive engineering and airfoil design in the domain of aerodynamics. The models show good accuracy in their predictions on a separate set of test geometries after training, with almost instant prediction speeds, as compared to O(hour) for the high-fidelity simulations required otherwise.
comment: 7 pages, 5 figures, 2022 IEEE Symposium Series on Computational Intelligence
☆ $\texttt{DoCoFL}$: Downlink Compression for Cross-Device Federated Learning
Many compression techniques have been proposed to reduce the communication overhead of Federated Learning training procedures. However, these are typically designed for compressing model updates, which are expected to decay throughout training. As a result, such methods are inapplicable to downlink (i.e., from the parameter server to clients) compression in the cross-device setting, where heterogeneous clients $\textit{may appear only once}$ during training and thus must download the model parameters. In this paper, we propose a new framework ($\texttt{DoCoFL}$) for downlink compression in the cross-device federated learning setting. Importantly, $\texttt{DoCoFL}$ can be seamlessly combined with many uplink compression schemes, rendering it suitable for bi-directional compression. Through extensive evaluation, we demonstrate that $\texttt{DoCoFL}$ offers significant bi-directional bandwidth reduction while achieving competitive accuracy to that of $\texttt{FedAvg}$ without compression.
☆ Analyzing Leakage of Personally Identifiable Information in Language Models
Language Models (LMs) have been shown to leak information about training data through sentence-level membership inference and reconstruction attacks. Understanding the risk of LMs leaking Personally Identifiable Information (PII) has received less attention, which can be attributed to the false assumption that dataset curation techniques such as scrubbing are sufficient to prevent PII leakage. Scrubbing techniques reduce but do not prevent the risk of PII leakage: in practice scrubbing is imperfect and must balance the trade-off between minimizing disclosure and preserving the utility of the dataset. On the other hand, it is unclear to which extent algorithmic defenses such as differential privacy, designed to guarantee sentence- or user-level privacy, prevent PII disclosure. In this work, we propose (i) a taxonomy of PII leakage in LMs, (ii) metrics to quantify PII leakage, and (iii) attacks showing that PII leakage is a threat in practice. Our taxonomy provides rigorous game-based definitions for PII leakage via black-box extraction, inference, and reconstruction attacks with only API access to an LM. We empirically evaluate attacks against GPT-2 models fine-tuned on three domains: case law, health care, and e-mails. Our main contributions are (i) novel attacks that can extract up to 10 times more PII sequences as existing attacks, (ii) showing that sentence-level differential privacy reduces the risk of PII disclosure but still leaks about 3% of PII sequences, and (iii) a subtle connection between record-level membership inference and PII reconstruction.
☆ Experimental observation on a low-rank tensor model for eigenvalue problems
Here we utilize a low-rank tensor model (LTM) as a function approximator, combined with the gradient descent method, to solve eigenvalue problems including the Laplacian operator and the harmonic oscillator. Experimental results show the superiority of the polynomial-based low-rank tensor model (PLTM) compared to the tensor neural network (TNN). We also test such low-rank architectures for the classification problem on the MNIST dataset.
☆ Effectiveness of Moving Target Defenses for Adversarial Attacks in ML-based Malware Detection
Several moving target defenses (MTDs) to counter adversarial ML attacks have been proposed in recent years. MTDs claim to increase the difficulty for the attacker in conducting attacks by regularly changing certain elements of the defense, such as cycling through configurations. To examine these claims, we study for the first time the effectiveness of several recent MTDs for adversarial ML attacks applied to the malware detection domain. Under different threat models, we show that transferability and query attack strategies can achieve high levels of evasion against these defenses through existing and novel attack strategies across Android and Windows. We also show that fingerprinting and reconnaissance are possible and demonstrate how attackers may obtain critical defense hyperparameters as well as information about how predictions are produced. Based on our findings, we present key recommendations for future work on the development of effective MTDs for adversarial attacks in ML-based malware detection.
☆ Distillation Policy Optimization
On-policy algorithms are supposed to be stable, however, sample-intensive yet. Off-policy algorithms utilizing past experiences are deemed to be sample-efficient, nevertheless, unstable in general. Can we design an algorithm that can employ the off-policy data, while exploit the stable learning by sailing along the course of the on-policy walkway? In this paper, we present an actor-critic learning framework that borrows the distributional perspective of interest to evaluate, and cross-breeds two sources of the data for policy improvement, which enables fast learning and can be applied to a wide class of algorithms. In its backbone, the variance reduction mechanisms, such as unified advantage estimator (UAE), that extends generalized advantage estimator (GAE) to be applicable on any state-dependent baseline, and a learned baseline, that is competent to stabilize the policy gradient, are firstly put forward to not merely be a bridge to the action-value function but also distill the advantageous learning signal. Lastly, it is empirically shown that our method improves sample efficiency and interpolates different levels well. Being of an organic whole, its mixture places more inspiration to the algorithm design.
☆ Off-the-Grid MARL: a Framework for Dataset Generation with Baselines for Cooperative Offline Multi-Agent Reinforcement Learning
Being able to harness the power of large, static datasets for developing autonomous multi-agent systems could unlock enormous value for real-world applications. Many important industrial systems are multi-agent in nature and are difficult to model using bespoke simulators. However, in industry, distributed system processes can often be recorded during operation, and large quantities of demonstrative data can be stored. Offline multi-agent reinforcement learning (MARL) provides a promising paradigm for building effective online controllers from static datasets. However, offline MARL is still in its infancy, and, therefore, lacks standardised benchmarks, baselines and evaluation protocols typically found in more mature subfields of RL. This deficiency makes it difficult for the community to sensibly measure progress. In this work, we aim to fill this gap by releasing \emph{off-the-grid MARL (OG-MARL)}: a framework for generating offline MARL datasets and algorithms. We release an initial set of datasets and baselines for cooperative offline MARL, created using the framework, along with a standardised evaluation protocol. Our datasets provide settings that are characteristic of real-world systems, including complex dynamics, non-stationarity, partial observability, suboptimality and sparse rewards, and are generated from popular online MARL benchmarks. We hope that OG-MARL will serve the community and help steer progress in offline MARL, while also providing an easy entry point for researchers new to the field.
☆ Towards Implementing Energy-aware Data-driven Intelligence for Smart Health Applications on Mobile Platforms
Recent breakthrough technological progressions of powerful mobile computing resources such as low-cost mobile GPUs along with cutting-edge, open-source software architectures have enabled high-performance deep learning on mobile platforms. These advancements have revolutionized the capabilities of today's mobile applications in different dimensions to perform data-driven intelligence locally, particularly for smart health applications. Unlike traditional machine learning (ML) architectures, modern on-device deep learning frameworks are proficient in utilizing computing resources in mobile platforms seamlessly, in terms of producing highly accurate results in less inference time. However, on the flip side, energy resources in a mobile device are typically limited. Hence, whenever a complex Deep Neural Network (DNN) architecture is fed into the on-device deep learning framework, while it achieves high prediction accuracy (and performance), it also urges huge energy demands during the runtime. Therefore, managing these resources efficiently within the spectrum of performance and energy efficiency is the newest challenge for any mobile application featuring data-driven intelligence beyond experimental evaluations. In this paper, first, we provide a timely review of recent advancements in on-device deep learning while empirically evaluating the performance metrics of current state-of-the-art ML architectures and conventional ML approaches with the emphasis given on energy characteristics by deploying them on a smart health application. With that, we are introducing a new framework through an energy-aware, adaptive model comprehension and realization (EAMCR) approach that can be utilized to make more robust and efficient inference decisions based on the available computing/energy resources in the mobile device during the runtime.
☆ Iterative Deepening Hyperband
Hyperparameter optimization (HPO) is concerned with the automated search for the most appropriate hyperparameter configuration (HPC) of a parameterized machine learning algorithm. A state-of-the-art HPO method is Hyperband, which, however, has its own parameters that influence its performance. One of these parameters, the maximal budget, is especially problematic: If chosen too small, the budget needs to be increased in hindsight and, as Hyperband is not incremental by design, the entire algorithm must be re-run. This is not only costly but also comes with a loss of valuable knowledge already accumulated. In this paper, we propose incremental variants of Hyperband that eliminate these drawbacks, and show that these variants satisfy theoretical guarantees qualitatively similar to those for the original Hyperband with the "right" budget. Moreover, we demonstrate their practical utility in experiments with benchmark data sets.
☆ Exploring Semantic Perturbations on Grover
With news and information being as easy to access as they currently are, it is more important than ever to ensure that people are not mislead by what they read. Recently, the rise of neural fake news (AI-generated fake news) and its demonstrated effectiveness at fooling humans has prompted the development of models to detect it. One such model is the Grover model, which can both detect neural fake news to prevent it, and generate it to demonstrate how a model could be misused to fool human readers. In this work we explore the Grover model's fake news detection capabilities by performing targeted attacks through perturbations on input news articles. Through this we test Grover's resilience to these adversarial attacks and expose some potential vulnerabilities which should be addressed in further iterations to ensure it can detect all types of fake news accurately.
comment: 15 pages, 12 figures, 1 table, capstone research in machine learning
☆ Learning Prototype Classifiers for Long-Tailed Recognition
The problem of long-tailed recognition (LTR) has received attention in recent years due to the fundamental power-law distribution of objects in the real-world. Most recent works in LTR use softmax classifiers that have a tendency to correlate classifier norm with the amount of training data for a given class. On the other hand, Prototype classifiers do not suffer from this shortcoming and can deliver promising results simply using Nearest-Class-Mean (NCM), a special case where prototypes are empirical centroids. However, the potential of Prototype classifiers as an alternative to softmax in LTR is relatively underexplored. In this work, we propose Prototype classifiers, which jointly learn prototypes that minimize average cross-entropy loss based on probability scores from distances to prototypes. We theoretically analyze the properties of Euclidean distance based prototype classifiers that leads to stable gradient-based optimization which is robust to outliers. We further enhance Prototype classifiers by learning channel-dependent temperature parameters to enable independent distance scales along each channel. Our analysis shows that prototypes learned by Prototype classifiers are better separated than empirical centroids. Results on four long-tailed recognition benchmarks show that Prototype classifier outperforms or is comparable to the state-of-the-art methods.
comment: 11 pages, 4 figures and 9 tables
☆ Equivariant Message Passing Neural Network for Crystal Material Discovery
Automatic material discovery with desired properties is a fundamental challenge for material sciences. Considerable attention has recently been devoted to generating stable crystal structures. While existing work has shown impressive success on supervised tasks such as property prediction, the progress on unsupervised tasks such as material generation is still hampered by the limited extent to which the equivalent geometric representations of the same crystal are considered. To address this challenge, we propose EMPNN a periodic equivariant message-passing neural network that learns crystal lattice deformation in an unsupervised fashion. Our model equivalently acts on lattice according to the deformation action that must be performed, making it suitable for crystal generation, relaxation and optimisation. We present experimental evaluations that demonstrate the effectiveness of our approach.
Conditional Flow Matching: Simulation-Free Dynamic Optimal Transport
Continuous normalizing flows (CNFs) are an attractive generative modeling technique, but they have thus far been held back by limitations in their simulation-based maximum likelihood training. In this paper, we introduce a new technique called conditional flow matching (CFM), a simulation-free training objective for CNFs. CFM features a stable regression objective like that used to train the stochastic flow in diffusion models but enjoys the efficient inference of deterministic flow models. In contrast to both diffusion models and prior CNF training algorithms, our CFM objective does not require the source distribution to be Gaussian or require evaluation of its density. Based on this new objective, we also introduce optimal transport CFM (OT-CFM), which creates simpler flows that are more stable to train and lead to faster inference, as evaluated in our experiments. Training CNFs with CFM improves results on a variety of conditional and unconditional generation tasks such as inferring single cell dynamics, unsupervised image translation, and Schr\"odinger bridge inference. Code is available at https://github.com/atong01/conditional-flow-matching .
☆ Improved Exact and Heuristic Algorithms for Maximum Weight Clique
We propose improved exact and heuristic algorithms for solving the maximum weight clique problem, a well-known problem in graph theory with many applications. Our algorithms interleave successful techniques from related work with novel data reduction rules that use local graph structure to identify and remove vertices and edges while retaining the optimal solution. We evaluate our algorithms on a range of synthetic and real-world graphs, and find that they outperform the current state of the art on most inputs. Our data reductions always produce smaller reduced graphs than existing data reductions alone. As a result, our exact algorithm, MWCRedu, finds solutions orders of magnitude faster on naturally weighted, medium-sized map labeling graphs and random hyperbolic graphs. Our heuristic algorithm, MWCPeel, outperforms its competitors on these instances, but is slightly less effective on extremely dense or large instances.
☆ Simplicity Bias in 1-Hidden Layer Neural Networks
Recent works have demonstrated that neural networks exhibit extreme simplicity bias(SB). That is, they learn only the simplest features to solve a task at hand, even in the presence of other, more robust but more complex features. Due to the lack of a general and rigorous definition of features, these works showcase SB on semi-synthetic datasets such as Color-MNIST, MNIST-CIFAR where defining features is relatively easier. In this work, we rigorously define as well as thoroughly establish SB for one hidden layer neural networks. More concretely, (i) we define SB as the network essentially being a function of a low dimensional projection of the inputs (ii) theoretically, we show that when the data is linearly separable, the network primarily depends on only the linearly separable ($1$-dimensional) subspace even in the presence of an arbitrarily large number of other, more complex features which could have led to a significantly more robust classifier, (iii) empirically, we show that models trained on real datasets such as Imagenette and Waterbirds-Landbirds indeed depend on a low dimensional projection of the inputs, thereby demonstrating SB on these datasets, iv) finally, we present a natural ensemble approach that encourages diversity in models by training successive models on features not used by earlier models, and demonstrate that it yields models that are significantly more robust to Gaussian noise.
☆ Width and Depth Limits Commute in Residual Networks
We show that taking the width and depth to infinity in a deep neural network with skip connections, when branches are scaled by $1/\sqrt{depth}$ (the only nontrivial scaling), result in the same covariance structure no matter how that limit is taken. This explains why the standard infinite-width-then-depth approach provides practical insights even for networks with depth of the same order as width. We also demonstrate that the pre-activations, in this case, have Gaussian distributions which has direct applications in Bayesian deep learning. We conduct extensive simulations that show an excellent match with our theoretical findings.
comment: 28 pages, 8 figures. arXiv admin note: text overlap with arXiv:2210.00688
☆ Deep Power Laws for Hyperparameter Optimization
Hyperparameter optimization is an important subfield of machine learning that focuses on tuning the hyperparameters of a chosen algorithm to achieve peak performance. Recently, there has been a stream of methods that tackle the issue of hyperparameter optimization, however, most of the methods do not exploit the scaling law property of learning curves. In this work, we propose Deep Power Laws (DPL), an ensemble of neural network models conditioned to yield predictions that follow a power-law scaling pattern. Our method dynamically decides which configurations to pause and train incrementally by making use of gray-box evaluations. We compare our method against 7 state-of-the-art competitors on 3 benchmarks related to tabular, image, and NLP datasets covering 57 diverse tasks. Our method achieves the best results across all benchmarks by obtaining the best any-time results compared to all competitors.
☆ Robust online active learning
In many industrial applications, obtaining labeled observations is not straightforward as it often requires the intervention of human experts or the use of expensive testing equipment. In these circumstances, active learning can be highly beneficial in suggesting the most informative data points to be used when fitting a model. Reducing the number of observations needed for model development alleviates both the computational burden required for training and the operational expenses related to labeling. Online active learning, in particular, is useful in high-volume production processes where the decision about the acquisition of the label for a data point needs to be taken within an extremely short time frame. However, despite the recent efforts to develop online active learning strategies, the behavior of these methods in the presence of outliers has not been thoroughly examined. In this work, we investigate the performance of online active linear regression in contaminated data streams. Our study shows that the currently available query strategies are prone to sample outliers, whose inclusion in the training set eventually degrades the predictive performance of the models. To address this issue, we propose a solution that bounds the search area of a conditional D-optimal algorithm and uses a robust estimator. Our approach strikes a balance between exploring unseen regions of the input space and protecting against outliers. Through numerical simulations, we show that the proposed method is effective in improving the performance of online active learning in the presence of outliers, thus expanding the potential applications of this powerful tool.
☆ Learning Choice Functions with Gaussian Processes
In consumer theory, ranking available objects by means of preference relations yields the most common description of individual choices. However, preference-based models assume that individuals: (1) give their preferences only between pairs of objects; (2) are always able to pick the best preferred object. In many situations, they may be instead choosing out of a set with more than two elements and, because of lack of information and/or incomparability (objects with contradictory characteristics), they may not able to select a single most preferred object. To address these situations, we need a choice-model which allows an individual to express a set-valued choice. Choice functions provide such a mathematical framework. We propose a Gaussian Process model to learn choice functions from choice-data. The proposed model assumes a multiple utility representation of a choice function based on the concept of Pareto rationalization, and derives a strategy to learn both the number and the values of these latent multiple utilities. Simulation experiments demonstrate that the proposed model outperforms the state-of-the-art methods.
☆ Deterministic equivalent and error universality of deep random features learning
This manuscript considers the problem of learning a random Gaussian network function using a fully connected network with frozen intermediate layers and trainable readout layer. This problem can be seen as a natural generalization of the widely studied random features model to deeper architectures. First, we prove Gaussian universality of the test error in a ridge regression setting where the learner and target networks share the same intermediate layers, and provide a sharp asymptotic formula for it. Establishing this result requires proving a deterministic equivalent for traces of the deep random features sample covariance matrices which can be of independent interest. Second, we conjecture the asymptotic Gaussian universality of the test error in the more general setting of arbitrary convex losses and generic learner/target architectures. We provide extensive numerical evidence for this conjecture, which requires the derivation of closed-form expressions for the layer-wise post-activation population covariances. In light of our results, we investigate the interplay between architecture design and implicit regularization.
☆ Delayed Feedback in Kernel Bandits
Black box optimisation of an unknown function from expensive and noisy evaluations is a ubiquitous problem in machine learning, academic research and industrial production. An abstraction of the problem can be formulated as a kernel based bandit problem (also known as Bayesian optimisation), where a learner aims at optimising a kernelized function through sequential noisy observations. The existing work predominantly assumes feedback is immediately available; an assumption which fails in many real world situations, including recommendation systems, clinical trials and hyperparameter tuning. We consider a kernel bandit problem under stochastically delayed feedback, and propose an algorithm with $\tilde{\mathcal{O}}(\sqrt{\Gamma_k(T)T}+\mathbb{E}[\tau])$ regret, where $T$ is the number of time steps, $\Gamma_k(T)$ is the maximum information gain of the kernel with $T$ observations, and $\tau$ is the delay random variable. This represents a significant improvement over the state of the art regret bound of $\tilde{\mathcal{O}}(\Gamma_k(T)\sqrt{T}+\mathbb{E}[\tau]\Gamma_k(T))$ reported in Verma et al. (2022). In particular, for very non-smooth kernels, the information gain grows almost linearly in time, trivializing the existing results. We also validate our theoretical results with simulations.
☆ Hierarchical Classification of Research Fields in the "Web of Science" Using Deep Learning
The scholarly publication space is growing steadily not just in numbers but also in complexity due to collaboration between individuals from within and across fields of research. This paper presents a hierarchical classification system that automatically categorizes a scholarly publication using its abstract into a three-tier hierarchical label set of fields (discipline-field-subfield). This system enables a holistic view about the interdependence of research activities in the mentioned hierarchical tiers in terms of knowledge production through articles and impact through citations. The classification system (44 disciplines - 738 fields - 1,501 subfields) utilizes and is able to cope with 160 million abstract snippets in Microsoft Academic Graph (Version 2018-05-17) using batch training in a modularized and distributed fashion to address and assess interdisciplinarity and inter-field classifications. In addition, we have explored multi-class classifications in both the single-label and multi-label settings. In total, we have conducted 3,140 experiments, in all models (Convolutional Neural Networks, Recurrent Neural Networks, Transformers), the classification accuracy is > 90% in 77.84% and 78.83% of the single-label and multi-label classifications, respectively. We examine the advantages of our classification by its ability to better align research texts and output with disciplines, to adequately classify them in an automated way, as well as to capture the degree of interdisciplinarity in a publication which enables downstream analytics such as field interdisciplinarity. This system (a set of pretrained models) can serve as a backbone to an interactive system of indexing scientific publications.
comment: Under review
☆ Short-term Prediction and Filtering of Solar Power Using State-Space Gaussian Processes NeurIPS 2022
Short-term forecasting of solar photovoltaic energy (PV) production is important for powerplant management. Ideally these forecasts are equipped with error bars, so that downstream decisions can account for uncertainty. To produce predictions with error bars in this setting, we consider Gaussian processes (GPs) for modelling and predicting solar photovoltaic energy production in the UK. A standard application of GP regression on the PV timeseries data is infeasible due to the large data size and non-Gaussianity of PV readings. However, this is made possible by leveraging recent advances in scalable GP inference, in particular, by using the state-space form of GPs, combined with modern variational inference techniques. The resulting model is not only scalable to large datasets but can also handle continuous data streams via Kalman filtering.
comment: Workshop paper submitted to "Tackling Climate Change with Machine Learning: workshop at NeurIPS 2022"
☆ Optimal Learning of Deep Random Networks of Extensive-width
We consider the problem of learning a target function corresponding to a deep, extensive-width, non-linear neural network with random Gaussian weights. We consider the asymptotic limit where the number of samples, the input dimension and the network width are proportionally large. We derive a closed-form expression for the Bayes-optimal test error, for regression and classification tasks. We contrast these Bayes-optimal errors with the test errors of ridge regression, kernel and random features regression. We find, in particular, that optimally regularized ridge regression, as well as kernel regression, achieve Bayes-optimal performances, while the logistic loss yields a near-optimal test error for classification. We further show numerically that when the number of samples grows faster than the dimension, ridge and kernel methods become suboptimal, while neural networks achieve test error close to zero from quadratically many samples.
☆ HOAX: A Hyperparameter Optimization Algorithm Explorer for Neural Networks
Computational chemistry has become an important tool to predict and understand molecular properties and reactions. Even though recent years have seen a significant growth in new algorithms and computational methods that speed up quantum chemical calculations, the bottleneck for trajectory-based methods to study photoinduced processes is still the huge number of electronic structure calculations. In this work, we present an innovative solution, in which the amount of electronic structure calculations is drastically reduced, by employing machine learning algorithms and methods borrowed from the realm of artificial intelligence. However, applying these algorithms effectively requires finding optimal hyperparameters, which remains a challenge itself. Here we present an automated user-friendly framework, HOAX, to perform the hyperparameter optimization for neural networks, which bypasses the need for a lengthy manual process. The neural network generated potential energy surfaces (PESs) reduces the computational costs compared to the ab initio-based PESs. We perform a comparative investigation on the performance of different hyperparameter optimiziation algorithms, namely grid search, simulated annealing, genetic algorithm, and bayesian optimizer in finding the optimal hyperparameters necessary for constructing the well-performing neural network in order to fit the PESs of small organic molecules. Our results show that this automated toolkit not only facilitate a straightforward way to perform the hyperparameter optimization but also the resulting neural networks-based generated PESs are in reasonable agreement with the ab initio-based PESs.
comment: 18 pages
☆ Simple yet Effective Gradient-Free Graph Convolutional Networks
Linearized Graph Neural Networks (GNNs) have attracted great attention in recent years for graph representation learning. Compared with nonlinear Graph Neural Network (GNN) models, linearized GNNs are much more time-efficient and can achieve comparable performances on typical downstream tasks such as node classification. Although some linearized GNN variants are purposely crafted to mitigate ``over-smoothing", empirical studies demonstrate that they still somehow suffer from this issue. In this paper, we instead relate over-smoothing with the vanishing gradient phenomenon and craft a gradient-free training framework to achieve more efficient and effective linearized GNNs which can significantly overcome over-smoothing and enhance the generalization of the model. The experimental results demonstrate that our methods achieve better and more stable performances on node classification tasks with varying depths and cost much less training time.
☆ How to select predictive models for causal inference?
Predictive models -- as with machine learning -- can underpin causal inference, to estimate the effects of an intervention at the population or individual level. This opens the door to a plethora of models, useful to match the increasing complexity of health data, but also the Pandora box of model selection: which of these models yield the most valid causal estimates? Classic machine-learning cross-validation procedures are not directly applicable. Indeed, an appropriate selection procedure for causal inference should equally weight both outcome errors for each individual, treated or not treated, whereas one outcome may be seldom observed for a sub-population. We study how more elaborate risks benefit causal model selection. We show theoretically that simple risks are brittle to weak overlap between treated and non-treated individuals as well as to heterogeneous errors between populations. Rather a more elaborate metric, the R-risk appears as a proxy of the oracle error on causal estimates, observable at the cost of an overlap re-weighting. As the R-risk is defined not only from model predictions but also by using the conditional mean outcome and the treatment probability, using it for model selection requires adapting cross validation. Extensive experiments show that the resulting procedure gives the best causal model selection.
comment: 31 pages
☆ Test-Time Amendment with a Coarse Classifier for Fine-Grained Classification
We investigate the problem of reducing mistake severity for fine-grained classification. Fine-grained classification can be challenging, mainly due to the requirement of knowledge or domain expertise for accurate annotation. However, humans are particularly adept at performing coarse classification as it requires relatively low levels of expertise. To this end, we present a novel approach for Post-Hoc Correction called Hierarchical Ensembles (HiE) that utilizes label hierarchy to improve the performance of fine-grained classification at test-time using the coarse-grained predictions. By only requiring the parents of leaf nodes, our method significantly reduces avg. mistake severity while improving top-1 accuracy on the iNaturalist-19 and tieredImageNet-H datasets, achieving a new state-of-the-art on both benchmarks. We also investigate the efficacy of our approach in the semi-supervised setting. Our approach brings notable gains in top-1 accuracy while significantly decreasing the severity of mistakes as training data decreases for the fine-grained classes. The simplicity and post-hoc nature of HiE render it practical to be used with any off-the-shelf trained model to improve its predictions further.
comment: 8 pages, 2 figures, 3 tables
☆ Bandit Convex Optimisation Revisited: FTRL Achieves $\tilde{O}(t^{1/2})$ Regret
We show that a kernel estimator using multiple function evaluations can be easily converted into a sampling-based bandit estimator with expectation equal to the original kernel estimate. Plugging such a bandit estimator into the standard FTRL algorithm yields a bandit convex optimisation algorithm that achieves $\tilde{O}(t^{1/2})$ regret against adversarial time-varying convex loss functions.
☆ Towards Label-Efficient Incremental Learning: A Survey
The current dominant paradigm when building a machine learning model is to iterate over a dataset over and over until convergence. Such an approach is non-incremental, as it assumes access to all images of all categories at once. However, for many applications, non-incremental learning is unrealistic. To that end, researchers study incremental learning, where a learner is required to adapt to an incoming stream of data with a varying distribution while preventing forgetting of past knowledge. Significant progress has been made, however, the vast majority of works focus on the fully supervised setting, making these algorithms label-hungry thus limiting their real-life deployment. To that end, in this paper, we make the first attempt to survey recently growing interest in label-efficient incremental learning. We identify three subdivisions, namely semi-, few-shot- and self-supervised learning to reduce labeling efforts. Finally, we identify novel directions that can further enhance label-efficiency and improve incremental learning scalability. Project website: {https://github.com/kilickaya/label-efficient-il.
☆ Anderson Acceleration For Bioinformatics-Based Machine Learning
Anderson acceleration (AA) is a well-known method for accelerating the convergence of iterative algorithms, with applications in various fields including deep learning and optimization. Despite its popularity in these areas, the effectiveness of AA in classical machine learning classifiers has not been thoroughly studied. Tabular data, in particular, presents a unique challenge for deep learning models, and classical machine learning models are known to perform better in these scenarios. However, the convergence analysis of these models has received limited attention. To address this gap in research, we implement a support vector machine (SVM) classifier variant that incorporates AA to speed up convergence. We evaluate the performance of our SVM with and without Anderson acceleration on several datasets from the biology domain and demonstrate that the use of AA significantly improves convergence and reduces the training loss as the number of iterations increases. Our findings provide a promising perspective on the potential of Anderson acceleration in the training of simple machine learning classifiers and underscore the importance of further research in this area. By showing the effectiveness of AA in this setting, we aim to inspire more studies that explore the applications of AA in classical machine learning.
☆ Predicting CSI Sequences With Attention-Based Neural Networks
In this work, we consider the problem of multi-step channel prediction in wireless communication systems. In existing works, autoregressive (AR) models are either replaced or combined with feed-forward neural networks(NNs) or, alternatively, with recurrent neural networks (RNNs). This paper explores the possibility of using sequence-to-sequence (Seq2Seq) and transformer neural network (TNN) models for channel state information (CSI) prediction. Simulation results show that both, Seq2Seq and TNNs, represent an appealing alternative to RNNs and feed-forward NNs in the context of CSI prediction. Additionally, the TNN with a few adaptations can extrapolate better than other models to CSI sequences that are either shorter or longer than the ones the model saw during training.
comment: Submitted to IEEE for publication
☆ Deep learning for $ψ$-weakly dependent processes
In this paper, we perform deep neural networks for learning $\psi$-weakly dependent processes. Such weak-dependence property includes a class of weak dependence conditions such as mixing, association,$\cdots$ and the setting considered here covers many commonly used situations such as: regression estimation, time series prediction, time series classification,$\cdots$ The consistency of the empirical risk minimization algorithm in the class of deep neural networks predictors is established. We achieve the generalization bound and obtain a learning rate, which is less than $\mathcal{O}(n^{-1/\alpha})$, for all $\alpha > 2 $. Applications to binary time series classification and prediction in affine causal models with exogenous covariates are carried out. Some simulation results are provided, as well as an application to the US recession data.
☆ Learning Functional Transduction
Research in Machine Learning has polarized into two general regression approaches: Transductive methods derive estimates directly from available data but are usually problem unspecific. Inductive methods can be much more particular, but generally require tuning and compute-intensive searches for solutions. In this work, we adopt a hybrid approach: We leverage the theory of Reproducing Kernel Banach Spaces (RKBS) and show that transductive principles can be induced through gradient descent to form efficient \textit{in-context} neural approximators. We apply this approach to RKBS of function-valued operators and show that once trained, our \textit{Transducer} model can capture on-the-fly relationships between infinite-dimensional input and output functions, given a few example pairs, and return new function estimates. We demonstrate the benefit of our transductive approach to model complex physical systems influenced by varying external factors with little data at a fraction of the usual deep learning training computation cost for partial differential equations and climate modeling applications.
☆ Development of deep biological ages aware of morbidity and mortality based on unsupervised and semi-supervised deep learning approaches
Background: While deep learning technology, which has the capability of obtaining latent representations based on large-scale data, can be a potential solution for the discovery of a novel aging biomarker, existing deep learning methods for biological age estimation usually depend on chronological ages and lack of consideration of mortality and morbidity that are the most significant outcomes of aging. Methods: This paper proposes a novel deep learning model to learn latent representations of biological aging in regard to subjects' morbidity and mortality. The model utilizes health check-up data in addition to morbidity and mortality information to learn the complex relationships between aging and measured clinical attributes. Findings: The proposed model is evaluated on a large dataset of general populations compared with KDM and other learning-based models. Results demonstrate that biological ages obtained by the proposed model have superior discriminability of subjects' morbidity and mortality.
☆ Accelerated First-Order Optimization under Nonlinear Constraints
We exploit analogies between first-order algorithms for constrained optimization and non-smooth dynamical systems to design a new class of accelerated first-order algorithms for constrained optimization. Unlike Frank-Wolfe or projected gradients, these algorithms avoid optimization over the entire feasible set at each iteration. We prove convergence to stationary points even in a nonconvex setting and we derive rates for the convex setting. An important property of these algorithms is that constraints are expressed in terms of velocities instead of positions, which naturally leads to sparse, local and convex approximations of the feasible set (even if the feasible set is nonconvex). Thus, the complexity tends to grow mildly in the number of decision variables and in the number of constraints, which makes the algorithms suitable for machine learning applications. We apply our algorithms to a compressed sensing and a sparse regression problem, showing that we can treat nonconvex $\ell^p$ constraints ($p<1$) efficiently, while recovering state-of-the-art performance for $p=1$.
comment: 25 pages, 5 figures
☆ Learning from Stochastic Labels
Annotating multi-class instances is a crucial task in the field of machine learning. Unfortunately, identifying the correct class label from a long sequence of candidate labels is time-consuming and laborious. To alleviate this problem, we design a novel labeling mechanism called stochastic label. In this setting, stochastic label includes two cases: 1) identify a correct class label from a small number of randomly given labels; 2) annotate the instance with None label when given labels do not contain correct class label. In this paper, we propose a novel suitable approach to learn from these stochastic labels. We obtain an unbiased estimator that utilizes less supervised information in stochastic labels to train a multi-class classifier. Additionally, it is theoretically justifiable by deriving the estimation error bound of the proposed method. Finally, we conduct extensive experiments on widely-used benchmark datasets to validate the superiority of our method by comparing it with existing state-of-the-art methods.
☆ The geometry of hidden representations of large transformer models
Large transformers are powerful architectures for self-supervised analysis of data of various nature, ranging from protein sequences to text to images. In these models, the data representation in the hidden layers live in the same space, and the semantic structure of the dataset emerges by a sequence of functionally identical transformations between one representation and the next. We here characterize the geometric and statistical properties of these representations, focusing on the evolution of such proprieties across the layers. By analyzing geometric properties such as the intrinsic dimension (ID) and the neighbor composition we find that the representations evolve in a strikingly similar manner in transformers trained on protein language tasks and image reconstruction tasks. In the first layers, the data manifold expands, becoming high-dimensional, and then it contracts significantly in the intermediate layers. In the last part of the model, the ID remains approximately constant or forms a second shallow peak. We show that the semantic complexity of the dataset emerges at the end of the first peak. This phenomenon can be observed across many models trained on diverse datasets. Based on these observations, we suggest using the ID profile as an unsupervised proxy to identify the layers which are more suitable for downstream learning tasks.
☆ A Survey of Methods, Challenges and Perspectives in Causality
The Causality field aims to find systematic methods for uncovering cause-effect relationships. Such methods can find applications in many research fields, justifying a great interest in this domain. Machine Learning models have shown success in a large variety of tasks by extracting correlation patterns from high-dimensional data but still struggle when generalizing out of their initial distribution. As causal engines aim to learn mechanisms that are independent from a data distribution, combining Machine Learning with Causality has the potential to bring benefits to the two fields. In our work, we motivate this assumption and provide applications. We first perform an extensive overview of the theories and methods for Causality from different perspectives. We then provide a deeper look at the connections between Causality and Machine Learning and describe the challenges met by the two domains. We show the early attempts to bring the fields together and the possible perspectives for the future. We finish by providing a large variety of applications for techniques from Causality.
comment: 42 pages, 35 pages for the main paper and 7 pages for the supplement, 10 figures, submitted to ACM Computing Surveys
☆ Jointist: Simultaneous Improvement of Multi-instrument Transcription and Music Source Separation via Joint Training
In this paper, we introduce Jointist, an instrument-aware multi-instrument framework that is capable of transcribing, recognizing, and separating multiple musical instruments from an audio clip. Jointist consists of an instrument recognition module that conditions the other two modules: a transcription module that outputs instrument-specific piano rolls, and a source separation module that utilizes instrument information and transcription results. The joint training of the transcription and source separation modules serves to improve the performance of both tasks. The instrument module is optional and can be directly controlled by human users. This makes Jointist a flexible user-controllable framework. Our challenging problem formulation makes the model highly useful in the real world given that modern popular music typically consists of multiple instruments. Its novelty, however, necessitates a new perspective on how to evaluate such a model. In our experiments, we assess the proposed model from various aspects, providing a new evaluation perspective for multi-instrument transcription. Our subjective listening study shows that Jointist achieves state-of-the-art performance on popular music, outperforming existing multi-instrument transcription models such as MT3. %We also argue that transcription models can be used as a preprocessing module for other music analysis tasks. We conducted experiments on several downstream tasks and found that the proposed method improved transcription by more than 1 percentage points (ppt.), source separation by 5 SDR, downbeat detection by 1.8 ppt., chord recognition by 1.4 ppt., and key estimation by 1.4 ppt., when utilizing transcription results obtained from Jointist.
comment: arXiv admin note: text overlap with arXiv:2206.10805
☆ Selective Uncertainty Propagation in Offline RL
We study the finite-horizon offline reinforcement learning (RL) problem. Since actions at any state can affect next-state distributions, the related distributional shift challenges can make this problem far more statistically complex than offline policy learning for a finite sequence of stochastic contextual bandit environments. We formalize this insight by showing that the statistical hardness of offline RL instances can be measured by estimating the size of actions' impact on next-state distributions. Furthermore, this estimated impact allows us to propagate just enough value function uncertainty from future steps to avoid model exploitation, enabling us to develop algorithms that improve upon traditional pessimistic approaches for offline RL on statistically simple instances. Our approach is supported by theory and simulations.
☆ Learning Generalized Zero-Shot Learners for Open-Domain Image Geolocalization
Image geolocalization is the challenging task of predicting the geographic coordinates of origin for a given photo. It is an unsolved problem relying on the ability to combine visual clues with general knowledge about the world to make accurate predictions across geographies. We present $\href{https://huggingface.co/geolocal/StreetCLIP}{\text{StreetCLIP}}$, a robust, publicly available foundation model not only achieving state-of-the-art performance on multiple open-domain image geolocalization benchmarks but also doing so in a zero-shot setting, outperforming supervised models trained on more than 4 million images. Our method introduces a meta-learning approach for generalized zero-shot learning by pretraining CLIP from synthetic captions, grounding CLIP in a domain of choice. We show that our method effectively transfers CLIP's generalized zero-shot capabilities to the domain of image geolocalization, improving in-domain generalized zero-shot performance without finetuning StreetCLIP on a fixed set of classes.
☆ W2SAT: Learning to generate SAT instances from Weighted Literal Incidence Graphs
The Boolean Satisfiability (SAT) problem stands out as an attractive NP-complete problem in theoretic computer science and plays a central role in a broad spectrum of computing-related applications. Exploiting and tuning SAT solvers under numerous scenarios require massive high-quality industry-level SAT instances, which unfortunately are quite limited in the real world. To address the data insufficiency issue, in this paper, we propose W2SAT, a framework to generate SAT formulas by learning intrinsic structures and properties from given real-world/industrial instances in an implicit fashion. To this end, we introduce a novel SAT representation called Weighted Literal Incidence Graph (WLIG), which exhibits strong representation ability and generalizability against existing counterparts, and can be efficiently generated via a specialized learning-based graph generative model. Decoding from WLIGs into SAT problems is then modeled as finding overlapping cliques with a novel hill-climbing optimization method termed Optimal Weight Coverage (OWC). Experiments demonstrate the superiority of our WLIG-induced approach in terms of graph metrics, efficiency, and scalability in comparison to previous methods. Additionally, we discuss the limitations of graph-based SAT generation for real-world applications, especially when utilizing generated instances for SAT solver parameter-tuning, and pose some potential directions.
☆ CATFL: Certificateless Authentication-based Trustworthy Federated Learning for 6G Semantic Communications
Federated learning (FL) provides an emerging approach for collaboratively training semantic encoder/decoder models of semantic communication systems, without private user data leaving the devices. Most existing studies on trustworthy FL aim to eliminate data poisoning threats that are produced by malicious clients, but in many cases, eliminating model poisoning attacks brought by fake servers is also an important objective. In this paper, a certificateless authentication-based trustworthy federated learning (CATFL) framework is proposed, which mutually authenticates the identity of clients and server. In CATFL, each client verifies the server's signature information before accepting the delivered global model to ensure that the global model is not delivered by false servers. On the contrary, the server also verifies the server's signature information before accepting the delivered model updates to ensure that they are submitted by authorized clients. Compared to PKI-based methods, the CATFL can avoid too high certificate management overheads. Meanwhile, the anonymity of clients shields data poisoning attacks, while real-name registration may suffer from user-specific privacy leakage risks. Therefore, a pseudonym generation strategy is also presented in CATFL to achieve a trade-off between identity traceability and user anonymity, which is essential to conditionally prevent from user-specific privacy leakage. Theoretical security analysis and evaluation results validate the superiority of CATFL.
comment: Accepted by WCNC workshop 2023
☆ Internally Rewarded Reinforcement Learning
We study a class of reinforcement learning problems where the reward signals for policy learning are generated by a discriminator that is dependent on and jointly optimized with the policy. This interdependence between the policy and the discriminator leads to an unstable learning process because reward signals from an immature discriminator are noisy and impede policy learning, and conversely, an untrained policy impedes discriminator learning. We call this learning setting $\textit{Internally Rewarded Reinforcement Learning}$ (IRRL) as the reward is not provided directly by the environment but $\textit{internally}$ by the discriminator. In this paper, we formally formulate IRRL and present a class of problems that belong to IRRL. We theoretically derive and empirically analyze the effect of the reward function in IRRL and based on these analyses propose the clipped linear reward function. Experimental results show that the proposed reward function can consistently stabilize the training process by reducing the impact of reward noise, which leads to faster convergence and higher performance compared with baselines in diverse tasks.
comment: 24 pages
☆ Dictionary-based Manifold Learning
We propose a paradigm for interpretable Manifold Learning for scientific data analysis, whereby we parametrize a manifold with $d$ smooth functions from a scientist-provided dictionary of meaningful, domain-related functions. When such a parametrization exists, we provide an algorithm for finding it based on sparse non-linear regression in the manifold tangent bundle, bypassing more standard manifold learning algorithms. We also discuss conditions for the existence of such parameterizations in function space and for successful recovery from finite samples. We demonstrate our method with experimental results from a real scientific domain.
☆ Implicit Regularization Leads to Benign Overfitting for Sparse Linear Regression
In deep learning, often the training process finds an interpolator (a solution with 0 training loss), but the test loss is still low. This phenomenon, known as benign overfitting, is a major mystery that received a lot of recent attention. One common mechanism for benign overfitting is implicit regularization, where the training process leads to additional properties for the interpolator, often characterized by minimizing certain norms. However, even for a simple sparse linear regression problem $y = \beta^{*\top} x +\xi$ with sparse $\beta^*$, neither minimum $\ell_1$ or $\ell_2$ norm interpolator gives the optimal test loss. In this work, we give a different parametrization of the model which leads to a new implicit regularization effect that combines the benefit of $\ell_1$ and $\ell_2$ interpolators. We show that training our new model via gradient descent leads to an interpolator with near-optimal test loss. Our result is based on careful analysis of the training dynamics and provides another example of implicit regularization effect that goes beyond norm minimization.
☆ QLAB: Quadratic Loss Approximation-Based Optimal Learning Rate for Deep Learning
We propose a learning rate adaptation scheme, called QLAB, for descent optimizers. We derive QLAB by optimizing the quadratic approximation of the loss function and QLAB can be combined with any optimizer who can provide the descent update direction. The computation of an adaptive learning rate with QLAB requires only computing an extra loss function value. We theoretically prove the convergence of the descent optimizers with QLAB. We demonstrate the effectiveness of QLAB in a range of optimization problems by combining with conclusively stochastic gradient descent, stochastic gradient descent with momentum, and Adam. The performance is validated on multi-layer neural networks, CNN, VGG-Net, ResNet and ShuffleNet with two datasets, MNIST and CIFAR10.
☆ Quickest Change Detection for Unnormalized Statistical Models AISTATS 2023
Classical quickest change detection algorithms require modeling pre-change and post-change distributions. Such an approach may not be feasible for various machine learning models because of the complexity of computing the explicit distributions. Additionally, these methods may suffer from a lack of robustness to model mismatch and noise. This paper develops a new variant of the classical Cumulative Sum (CUSUM) algorithm for the quickest change detection. This variant is based on Fisher divergence and the Hyv\"arinen score and is called the Score-based CUSUM (SCUSUM) algorithm. The SCUSUM algorithm allows the applications of change detection for unnormalized statistical models, i.e., models for which the probability density function contains an unknown normalization constant. The asymptotic optimality of the proposed algorithm is investigated by deriving expressions for average detection delay and the mean running time to a false alarm. Numerical results are provided to demonstrate the performance of the proposed algorithm.
comment: A version of this paper has been accepted by the 26th International Conference on Artificial Intelligence and Statistics (AISTATS 2023)
☆ A Nearly-Optimal Bound for Fast Regression with $\ell_\infty$ Guarantee
Given a matrix $A\in \mathbb{R}^{n\times d}$ and a vector $b\in \mathbb{R}^n$, we consider the regression problem with $\ell_\infty$ guarantees: finding a vector $x'\in \mathbb{R}^d$ such that $ \|x'-x^*\|_\infty \leq \frac{\epsilon}{\sqrt{d}}\cdot \|Ax^*-b\|_2\cdot \|A^\dagger\|$ where $x^*=\arg\min_{x\in \mathbb{R}^d}\|Ax-b\|_2$. One popular approach for solving such $\ell_2$ regression problem is via sketching: picking a structured random matrix $S\in \mathbb{R}^{m\times n}$ with $m\ll n$ and $SA$ can be quickly computed, solve the ``sketched'' regression problem $\arg\min_{x\in \mathbb{R}^d} \|SAx-Sb\|_2$. In this paper, we show that in order to obtain such $\ell_\infty$ guarantee for $\ell_2$ regression, one has to use sketching matrices that are dense. To the best of our knowledge, this is the first user case in which dense sketching matrices are necessary. On the algorithmic side, we prove that there exists a distribution of dense sketching matrices with $m=\epsilon^{-2}d\log^3(n/\delta)$ such that solving the sketched regression problem gives the $\ell_\infty$ guarantee, with probability at least $1-\delta$. Moreover, the matrix $SA$ can be computed in time $O(nd\log n)$. Our row count is nearly-optimal up to logarithmic factors, and significantly improves the result in [Price, Song and Woodruff, ICALP'17], in which a super-linear in $d$ rows, $m=\Omega(\epsilon^{-2}d^{1+\gamma})$ for $\gamma=\Theta(\sqrt{\frac{\log\log n}{\log d}})$ is required. We also develop a novel analytical framework for $\ell_\infty$ guarantee regression that utilizes the Oblivious Coordinate-wise Embedding (OCE) property introduced in [Song and Yu, ICML'21]. Our analysis is arguably much simpler and more general than [Price, Song and Woodruff, ICALP'17], and it extends to dense sketches for tensor product of vectors.
comment: Abstract shortened to meet arxiv requirement
☆ TAP: Accelerating Large-Scale DNN Training Through Tensor Automatic Parallelisation
Model parallelism has become necessary to train large neural networks. However, finding a suitable model parallel schedule for an arbitrary neural network is a non-trivial task due to the exploding search space. In this work, we present a model parallelism framework TAP that automatically searches for the best data and tensor parallel schedules. Leveraging the key insight that a neural network can be represented as a directed acyclic graph, within which may only exist a limited set of frequent subgraphs, we design a graph pruning algorithm to fold the search space efficiently. TAP runs at sub-linear complexity concerning the neural network size. Experiments show that TAP is $20\times- 160\times$ faster than the state-of-the-art automatic parallelism framework, and the performance of its discovered schedules is competitive with the expert-engineered ones.
☆ Learning Cut Selection for Mixed-Integer Linear Programming via Hierarchical Sequence Model ICLR2023
Cutting planes (cuts) are important for solving mixed-integer linear programs (MILPs), which formulate a wide range of important real-world applications. Cut selection -- which aims to select a proper subset of the candidate cuts to improve the efficiency of solving MILPs -- heavily depends on (P1) which cuts should be preferred, and (P2) how many cuts should be selected. Although many modern MILP solvers tackle (P1)-(P2) by manually designed heuristics, machine learning offers a promising approach to learn more effective heuristics from MILPs collected from specific applications. However, many existing learning-based methods focus on learning which cuts should be preferred, neglecting the importance of learning the number of cuts that should be selected. Moreover, we observe from extensive empirical results that (P3) what order of selected cuts should be preferred has a significant impact on the efficiency of solving MILPs as well. To address this challenge, we propose a novel hierarchical sequence model (HEM) to learn cut selection policies via reinforcement learning. Specifically, HEM consists of a two-level model: (1) a higher-level model to learn the number of cuts that should be selected, (2) and a lower-level model -- that formulates the cut selection task as a sequence to sequence learning problem -- to learn policies selecting an ordered subset with the size determined by the higher-level model. To the best of our knowledge, HEM is the first method that can tackle (P1)-(P3) in cut selection simultaneously from a data-driven perspective. Experiments show that HEM significantly improves the efficiency of solving MILPs compared to human-designed and learning-based baselines on both synthetic and large-scale real-world MILPs, including MIPLIB 2017. Moreover, experiments demonstrate that HEM well generalizes to MILPs that are significantly larger than those seen during training.
comment: Accepted to ICLR2023
☆ The Parametric Stability of Well-separated Spherical Gaussian Mixtures
We quantify the parameter stability of a spherical Gaussian Mixture Model (sGMM) under small perturbations in distribution space. Namely, we derive the first explicit bound to show that for a mixture of spherical Gaussian $P$ (sGMM) in a pre-defined model class, all other sGMM close to $P$ in this model class in total variation distance has a small parameter distance to $P$. Further, this upper bound only depends on $P$. The motivation for this work lies in providing guarantees for fitting Gaussian mixtures; with this aim in mind, all the constants involved are well defined and distribution free conditions for fitting mixtures of spherical Gaussians. Our results tighten considerably the existing computable bounds, and asymptotically match the known sharp thresholds for this problem.
☆ Filtering Context Mitigates Scarcity and Selection Bias in Political Ideology Prediction
We propose a novel supervised learning approach for political ideology prediction (PIP) that is capable of predicting out-of-distribution inputs. This problem is motivated by the fact that manual data-labeling is expensive, while self-reported labels are often scarce and exhibit significant selection bias. We propose a novel statistical model that decomposes the document embeddings into a linear superposition of two vectors; a latent neutral \emph{context} vector independent of ideology, and a latent \emph{position} vector aligned with ideology. We train an end-to-end model that has intermediate contextual and positional vectors as outputs. At deployment time, our model predicts labels for input documents by exclusively leveraging the predicted positional vectors. On two benchmark datasets we show that our model is capable of outputting predictions even when trained with as little as 5\% biased data, and is significantly more accurate than the state-of-the-art. Through crowd-sourcing we validate the neutrality of contextual vectors, and show that context filtering results in ideological concentration, allowing for prediction on out-of-distribution examples.
☆ Bridging Physics-Informed Neural Networks with Reinforcement Learning: Hamilton-Jacobi-Bellman Proximal Policy Optimization (HJBPPO)
This paper introduces the Hamilton-Jacobi-Bellman Proximal Policy Optimization (HJBPPO) algorithm into reinforcement learning. The Hamilton-Jacobi-Bellman (HJB) equation is used in control theory to evaluate the optimality of the value function. Our work combines the HJB equation with reinforcement learning in continuous state and action spaces to improve the training of the value network. We treat the value network as a Physics-Informed Neural Network (PINN) to solve for the HJB equation by computing its derivatives with respect to its inputs exactly. The Proximal Policy Optimization (PPO)-Clipped algorithm is improvised with this implementation as it uses a value network to compute the objective function for its policy network. The HJBPPO algorithm shows an improved performance compared to PPO on the MuJoCo environments.
☆ Generative Adversarial Symmetry Discovery
Despite the success of equivariant neural networks in scientific applications, they require knowing the symmetry group a priori. However, it may be difficult to know the right symmetry to use as an inductive bias in practice and enforcing the wrong symmetry could hurt the performance. In this paper, we propose a framework, LieGAN, to automatically discover equivariances from a dataset using a paradigm akin to generative adversarial training. Specifically, a generator learns a group of transformations applied to the data, which preserves the original distribution and fools the discriminator. LieGAN represents symmetry as interpretable Lie algebra basis and can discover various symmetries such as rotation group $\mathrm{SO}(n)$ and restricted Lorentz group $\mathrm{SO}(1,3)^+$ in trajectory prediction and top quark tagging tasks. The learned symmetry can also be readily used in several existing equivariant neural networks to improve accuracy and generalization in prediction.
☆ SPIDE: A Purely Spike-based Method for Training Feedback Spiking Neural Networks
Spiking neural networks (SNNs) with event-based computation are promising brain-inspired models for energy-efficient applications on neuromorphic hardware. However, most supervised SNN training methods, such as conversion from artificial neural networks or direct training with surrogate gradients, require complex computation rather than spike-based operations of spiking neurons during training. In this paper, we study spike-based implicit differentiation on the equilibrium state (SPIDE) that extends the recently proposed training method, implicit differentiation on the equilibrium state (IDE), for supervised learning with purely spike-based computation, which demonstrates the potential for energy-efficient training of SNNs. Specifically, we introduce ternary spiking neuron couples and prove that implicit differentiation can be solved by spikes based on this design, so the whole training procedure, including both forward and backward passes, is made as event-driven spike computation, and weights are updated locally with two-stage average firing rates. Then we propose to modify the reset membrane potential to reduce the approximation error of spikes. With these key components, we can train SNNs with flexible structures in a small number of time steps and with firing sparsity during training, and the theoretical estimation of energy costs demonstrates the potential for high efficiency. Meanwhile, experiments show that even with these constraints, our trained models can still achieve competitive results on MNIST, CIFAR-10, CIFAR-100, and CIFAR10-DVS. Our code is available at https://github.com/pkuxmq/SPIDE-FSNN.
comment: Accepted by Neural Networks
☆ Efficient Scopeformer: Towards Scalable and Rich Feature Extraction for Intracranial Hemorrhage Detection
The quality and richness of feature maps extracted by convolution neural networks (CNNs) and vision Transformers (ViTs) directly relate to the robust model performance. In medical computer vision, these information-rich features are crucial for detecting rare cases within large datasets. This work presents the "Scopeformer," a novel multi-CNN-ViT model for intracranial hemorrhage classification in computed tomography (CT) images. The Scopeformer architecture is scalable and modular, which allows utilizing various CNN architectures as the backbone with diversified output features and pre-training strategies. We propose effective feature projection methods to reduce redundancies among CNN-generated features and to control the input size of ViTs. Extensive experiments with various Scopeformer models show that the model performance is proportional to the number of convolutional blocks employed in the feature extractor. Using multiple strategies, including diversifying the pre-training paradigms for CNNs, different pre-training datasets, and style transfer techniques, we demonstrate an overall improvement in the model performance at various computational budgets. Later, we propose smaller compute-efficient Scopeformer versions with three different types of input and output ViT configurations. Efficient Scopeformers use four different pre-trained CNN architectures as feature extractors to increase feature richness. Our best Efficient Scopeformer model achieved an accuracy of 96.94\% and a weighted logarithmic loss of 0.083 with an eight times reduction in the number of trainable parameters compared to the base Scopeformer. Another version of the Efficient Scopeformer model further reduced the parameter space by almost 17 times with negligible performance reduction. Hybrid CNNs and ViTs might provide the desired feature richness for developing accurate medical computer vision models
☆ Knowledge Distillation on Graphs: A Survey
Graph Neural Networks (GNNs) have attracted tremendous attention by demonstrating their capability to handle graph data. However, they are difficult to be deployed in resource-limited devices due to model sizes and scalability constraints imposed by the multi-hop data dependency. In addition, real-world graphs usually possess complex structural information and features. Therefore, to improve the applicability of GNNs and fully encode the complicated topological information, knowledge distillation on graphs (KDG) has been introduced to build a smaller yet effective model and exploit more knowledge from data, leading to model compression and performance improvement. Recently, KDG has achieved considerable progress with many studies proposed. In this survey, we systematically review these works. Specifically, we first introduce KDG challenges and bases, then categorize and summarize existing works of KDG by answering the following three questions: 1) what to distillate, 2) who to whom, and 3) how to distillate. Finally, we share our thoughts on future research directions.
☆ QCRS: Improve Randomized Smoothing using Quasi-Concave Optimization
Randomized smoothing is currently the state-of-the-art method that provides certified robustness for deep neural networks. However, it often cannot achieve an adequate certified region on real-world datasets. One way to obtain a larger certified region is to use an input-specific algorithm instead of using a fixed Gaussian filter for all data points. Several methods based on this idea have been proposed, but they either suffer from high computational costs or gain marginal improvement in certified radius. In this work, we show that by exploiting the quasiconvex problem structure, we can find the optimal certified radii for most data points with slight computational overhead. This observation leads to an efficient and effective input-specific randomized smoothing algorithm. We conduct extensive experiments and empirical analysis on Cifar10 and ImageNet. The results show that the proposed method significantly enhances the certified radii with low computational overhead.
comment: 10 pages
☆ Distributed Traffic Synthesis and Classification in Edge Networks: A Federated Self-supervised Learning Approach
With the rising demand for wireless services and increased awareness of the need for data protection, existing network traffic analysis and management architectures are facing unprecedented challenges in classifying and synthesizing the increasingly diverse services and applications. This paper proposes FS-GAN, a federated self-supervised learning framework to support automatic traffic analysis and synthesis over a large number of heterogeneous datasets. FS-GAN is composed of multiple distributed Generative Adversarial Networks (GANs), with a set of generators, each being designed to generate synthesized data samples following the distribution of an individual service traffic, and each discriminator being trained to differentiate the synthesized data samples and the real data samples of a local dataset. A federated learning-based framework is adopted to coordinate local model training processes of different GANs across different datasets. FS-GAN can classify data of unknown types of service and create synthetic samples that capture the traffic distribution of the unknown types. We prove that FS-GAN can minimize the Jensen-Shannon Divergence (JSD) between the distribution of real data across all the datasets and that of the synthesized data samples. FS-GAN also maximizes the JSD among the distributions of data samples created by different generators, resulting in each generator producing synthetic data samples that follow the same distribution as one particular service type. Extensive simulation results show that the classification accuracy of FS-GAN achieves over 20% improvement in average compared to the state-of-the-art clustering-based traffic analysis algorithms. FS-GAN also has the capability to synthesize highly complex mixtures of traffic types without requiring any human-labeled data samples.
comment: published as early access at IEEE Transactions on Mobile Computing, January 2023
☆ Electrode Selection for Noninvasive Fetal Electrocardiogram Extraction using Mutual Information Criteria
Blind source separation (BSS) techniques have revealed to be promising approaches for, among other, biomedical signal processing applications. Specifically, for the noninvasive extraction of fetal cardiac signals from maternal abdominal recordings, where conventional filtering schemes have failed to extract the complete fetal ECG components. From previous studies, it is now believed that a carefully selected array of electrodes well-placed over the abdomen of a pregnant woman contains the required `information' for BSS, to extract the complete fetal components. Based on this idea, in previous works array recording systems and sensor selection strategies based on the Mutual Information (MI) criterion have been developed. In this paper the previous works have been extended, by considering the 3-dimensional aspects of the cardiac electrical activity. The proposed method has been tested on simulated and real maternal abdominal recordings. The results show that the new sensor selection strategy together with the MI criterion, can be effectively used to select the channels containing the most `information' concerning the fetal ECG components from an array of 72 recordings. The method is hence believed to be useful for the selection of the most informative channels in online applications, considering the different fetal positions and movements.
☆ Gradient Descent in Neural Networks as Sequential Learning in RKBS
The study of Neural Tangent Kernels (NTKs) has provided much needed insight into convergence and generalization properties of neural networks in the over-parametrized (wide) limit by approximating the network using a first-order Taylor expansion with respect to its weights in the neighborhood of their initialization values. This allows neural network training to be analyzed from the perspective of reproducing kernel Hilbert spaces (RKHS), which is informative in the over-parametrized regime, but a poor approximation for narrower networks as the weights change more during training. Our goal is to extend beyond the limits of NTK toward a more general theory. We construct an exact power-series representation of the neural network in a finite neighborhood of the initial weights as an inner product of two feature maps, respectively from data and weight-step space, to feature space, allowing neural network training to be analyzed from the perspective of reproducing kernel {\em Banach} space (RKBS). We prove that, regardless of width, the training sequence produced by gradient descent can be exactly replicated by regularized sequential learning in RKBS. Using this, we present novel bound on uniform convergence where the iterations count and learning rate play a central role, giving new theoretical insight into neural network training.
☆ End-to-End Full-Atom Antibody Design
Antibody design is an essential yet challenging task in various domains like therapeutics and biology. There are two major defects in current learning-based methods: 1) tackling only a certain subtask of the whole antibody design pipeline, making them suboptimal or resource-intensive. 2) omitting either the framework regions or side chains, thus incapable of capturing the full-atom geometry. To address these pitfalls, we propose dynamic Multi-channel Equivariant grAph Network (dyMEAN), an end-to-end full-atom model for E(3)-equivariant antibody design given the epitope and the incomplete sequence of the antibody. Specifically, we first explore structural initialization as a knowledgeable guess of the antibody structure and then propose shadow paratope to bridge the epitope-antibody connections. Both 1D sequences and 3D structures are updated via an adaptive multi-channel equivariant encoder that is able to process protein residues of variable sizes when considering full atoms. Finally, the updated antibody is docked to the epitope via the alignment of the shadow paratope. Experiments on epitope-binding CDR-H3 design, complex structure prediction, and affinity optimization demonstrate the superiority of our end-to-end framework and full-atom modeling.
comment: preprint
☆ A Transaction Represented with Weighted Finite-State Transducers
Not all contracts are good, but all good contracts can be expressed as a finite-state transition system ("State-Transition Contracts"). Contracts that can be represented as State-Transition Contracts discretize fat-tailed risk to foreseeable, managed risk, define the boundary of relevant events governed by the relationship, and eliminate the potential of inconsistent contractual provisions. Additionally, State-Transition Contracts reap the substantial benefit of being able to be analyzed under the rules governing the science of the theory of computation. Simple State-Transition Contracts can be represented as discrete finite automata; more complicated State-Transition Contracts, such as those that have downstream effects on other agreements or complicated pathways of performance, benefit from representation as weighted finite-state transducers, with weights assigned as costs, penalties, or probabilities of transitions. This research paper (the "Research" or "Paper") presents a complex legal transaction represented as weighted finite-state transducers. Furthermore, we show that the mathematics/algorithms permitted by the algebraic structure of weighted finite-state transducers provides actionable, legal insight into the transaction.
comment: 2 figures, 3 tables, 2 appendices, Recognition Technologies, Inc. Technical Report
☆ Weight Prediction Boosts the Convergence of AdamW
In this paper, we introduce weight prediction into the AdamW optimizer to boost its convergence when training the deep neural network (DNN) models. In particular, ahead of each mini-batch training, we predict the future weights according to the update rule of AdamW and then apply the predicted future weights to do both forward pass and backward propagation. In this way, the AdamW optimizer always utilizes the gradients w.r.t. the future weights instead of current weights to update the DNN parameters, making the AdamW optimizer achieve better convergence. Our proposal is simple and straightforward to implement but effective in boosting the convergence of DNN training. We performed extensive experimental evaluations on image classification and language modeling tasks to verify the effectiveness of our proposal. The experimental results validate that our proposal can boost the convergence of AdamW and achieve better accuracy than AdamW when training the DNN models.
☆ Free Lunch for Domain Adversarial Training: Environment Label Smoothing ICLR 2023
A fundamental challenge for machine learning models is how to generalize learned models for out-of-distribution (OOD) data. Among various approaches, exploiting invariant features by Domain Adversarial Training (DAT) received widespread attention. Despite its success, we observe training instability from DAT, mostly due to over-confident domain discriminator and environment label noise. To address this issue, we proposed Environment Label Smoothing (ELS), which encourages the discriminator to output soft probability, which thus reduces the confidence of the discriminator and alleviates the impact of noisy environment labels. We demonstrate, both experimentally and theoretically, that ELS can improve training stability, local convergence, and robustness to noisy environment labels. By incorporating ELS with DAT methods, we are able to yield state-of-art results on a wide range of domain generalization/adaptation tasks, particularly when the environment labels are highly noisy.
comment: ICLR 2023, 38 pages, 8 figures, 18 tables
☆ $\rm A^2Q$: Aggregation-Aware Quantization for Graph Neural Networks ICLR2023
As graph data size increases, the vast latency and memory consumption during inference pose a significant challenge to the real-world deployment of Graph Neural Networks (GNNs). While quantization is a powerful approach to reducing GNNs complexity, most previous works on GNNs quantization fail to exploit the unique characteristics of GNNs, suffering from severe accuracy degradation. Through an in-depth analysis of the topology of GNNs, we observe that the topology of the graph leads to significant differences between nodes, and most of the nodes in a graph appear to have a small aggregation value. Motivated by this, in this paper, we propose the Aggregation-Aware mixed-precision Quantization ($\rm A^2Q$) for GNNs, where an appropriate bitwidth is automatically learned and assigned to each node in the graph. To mitigate the vanishing gradient problem caused by sparse connections between nodes, we propose a Local Gradient method to serve the quantization error of the node features as the supervision during training. We also develop a Nearest Neighbor Strategy to deal with the generalization on unseen graphs. Extensive experiments on eight public node-level and graph-level datasets demonstrate the generality and robustness of our proposed method. Compared to the FP32 models, our method can achieve up to a 18.6x (i.e., 1.70bit) compression ratio with negligible accuracy degradation. Morever, compared to the state-of-the-art quantization method, our method can achieve up to 11.4\% and 9.5\% accuracy improvements on the node-level and graph-level tasks, respectively, and up to 2x speedup on a dedicated hardware accelerator.
comment: Accepted by ICLR2023
☆ Density peak clustering using tensor network
Tensor networks, which have been traditionally used to simulate many-body physics, have recently gained significant attention in the field of machine learning due to their powerful representation capabilities. In this work, we propose a density-based clustering algorithm inspired by tensor networks. We encode classical data into tensor network states on an extended Hilbert space and train the tensor network states to capture the features of the clusters. Here, we define density and related concepts in terms of fidelity, rather than using a classical distance measure. We evaluate the performance of our algorithm on six synthetic data sets, four real world data sets, and three commonly used computer vision data sets. The results demonstrate that our method provides state-of-the-art performance on several synthetic data sets and real world data sets, even when the number of clusters is unknown. Additionally, our algorithm performs competitively with state-of-the-art algorithms on the MNIST, USPS, and Fashion-MNIST image data sets. These findings reveal the great potential of tensor networks for machine learning applications.
☆ Deep Learning Approach to Predict Hemorrhage in Moyamoya Disease
Objective: Reliable tools to predict moyamoya disease (MMD) patients at risk for hemorrhage could have significant value. The aim of this paper is to develop three machine learning classification algorithms to predict hemorrhage in moyamoya disease. Methods: Clinical data of consecutive MMD patients who were admitted to our hospital between 2009 and 2015 were reviewed. Demographics, clinical, radiographic data were analyzed to develop artificial neural network (ANN), support vector machine (SVM), and random forest models. Results: We extracted 33 parameters, including 11 demographic and 22 radiographic features as input for model development. Of all compared classification results, ANN achieved the highest overall accuracy of 75.7% (95% CI, 68.6%-82.8%), followed by SVM with 69.2% (95% CI, 56.9%-81.5%) and random forest with 70.0% (95% CI, 57.0%-83.0%). Conclusions: The proposed ANN framework can be a potential effective tool to predict the possibility of hemorrhage among adult MMD patients based on clinical information and radiographic features.
☆ Program Generation from Diverse Video Demonstrations
The ability to use inductive reasoning to extract general rules from multiple observations is a vital indicator of intelligence. As humans, we use this ability to not only interpret the world around us, but also to predict the outcomes of the various interactions we experience. Generalising over multiple observations is a task that has historically presented difficulties for machines to grasp, especially when requiring computer vision. In this paper, we propose a model that can extract general rules from video demonstrations by simultaneously performing summarisation and translation. Our approach differs from prior works by framing the problem as a multi-sequence-to-sequence task, wherein summarisation is learnt by the model. This allows our model to utilise edge cases that would otherwise be suppressed or discarded by traditional summarisation techniques. Additionally, we show that our approach can handle noisy specifications without the need for additional filtering methods. We evaluate our model by synthesising programs from video demonstrations in the Vizdoom environment achieving state-of-the-art results with a relative increase of 11.75% program accuracy on prior works
☆ Active Uncertainty Reduction for Safe and Efficient Interaction Planning: A Shielding-Aware Dual Control Approach
The ability to accurately predict the opponent's behavior is central to the safety and efficiency of robotic systems in interactive settings, such as human-robot interaction and multi-robot teaming tasks. Unfortunately, robots often lack access to key information on which these predictions may hinge, such as opponent's goals, attention, and willingness to cooperate. Dual control theory addresses this challenge by treating unknown parameters of a predictive model as hidden states and inferring their values at runtime using information gathered during system operation. While able to optimally and automatically trade off exploration and exploitation, dual control is computationally intractable for general interactive motion planning. In this paper, we present a novel algorithmic approach to enable active uncertainty reduction for interactive motion planning based on the implicit dual control paradigm. Our approach relies on sampling-based approximation of stochastic dynamic programming, leading to a model predictive control problem. The resulting policy is shown to preserve the dual control effect for a broad class of predictive models with both continuous and categorical uncertainty. To ensure the safe operation of the interacting agents, we leverage a supervisory control scheme, oftentimes referred to as ``shielding'', which overrides the ego agent's dual control policy with a safety fallback strategy when a safety-critical event is imminent. We then augment the dual control framework with an improved variant of the recently proposed shielding-aware robust planning scheme, which proactively balances the nominal planning performance with the risk of high-cost emergency maneuvers triggered by low-probability opponent's behaviors. We demonstrate the efficacy of our approach with both simulated driving examples and hardware experiments using 1/10 scale autonomous vehicles.
comment: arXiv admin note: text overlap with arXiv:2202.07720
☆ Diffusion Models for High-Resolution Solar Forecasts
Forecasting future weather and climate is inherently difficult. Machine learning offers new approaches to increase the accuracy and computational efficiency of forecasts, but current methods are unable to accurately model uncertainty in high-dimensional predictions. Score-based diffusion models offer a new approach to modeling probability distributions over many dependent variables, and in this work, we demonstrate how they provide probabilistic forecasts of weather and climate variables at unprecedented resolution, speed, and accuracy. We apply the technique to day-ahead solar irradiance forecasts by generating many samples from a diffusion model trained to super-resolve coarse-resolution numerical weather predictions to high-resolution weather satellite observations.
☆ Detection of Tomato Ripening Stages using Yolov3-tiny
One of the most important agricultural products in Mexico is the tomato (Solanum lycopersicum), which occupies the 4th place national most produced product . Therefore, it is necessary to improve its production, building automatic detection system that detect, classify an keep tacks of the fruits is one way to archieve it. So, in this paper, we address the design of a computer vision system to detect tomatoes at different ripening stages. To solve the problem, we use a neural network-based model for tomato classification and detection. Specifically, we use the YOLOv3-tiny model because it is one of the lightest current deep neural networks. To train it, we perform two grid searches testing several combinations of hyperparameters. Our experiments showed an f1-score of 90.0% in the localization and classification of ripening stages in a custom dataset.
☆ FLSTRA: Federated Learning in Stratosphere
We propose a federated learning (FL) in stratosphere (FLSTRA) system, where a high altitude platform station (HAPS) felicitates a large number of terrestrial clients to collaboratively learn a global model without sharing the training data. FLSTRA overcomes the challenges faced by FL in terrestrial networks, such as slow convergence and high communication delay due to limited client participation and multi-hop communications. HAPS leverages its altitude and size to allow the participation of more clients with line-of-sight (LoS) links and the placement of a powerful server. However, handling many clients at once introduces computing and transmission delays. Thus, we aim to obtain a delay-accuracy trade-off for FLSTRA. Specifically, we first develop a joint client selection and resource allocation algorithm for uplink and downlink to minimize the FL delay subject to the energy and quality-of-service (QoS) constraints. Second, we propose a communication and computation resource-aware (CCRA-FL) algorithm to achieve the target FL accuracy while deriving an upper bound for its convergence rate. The formulated problem is non-convex; thus, we propose an iterative algorithm to solve it. Simulation results demonstrate the effectiveness of the proposed FLSTRA system, compared to terrestrial benchmarks, in terms of FL delay and accuracy.
comment: Submitted to IEEE for possible publication
☆ Local transfer learning from one data space to another
A fundamental problem in manifold learning is to approximate a functional relationship in a data chosen randomly from a probability distribution supported on a low dimensional sub-manifold of a high dimensional ambient Euclidean space. The manifold is essentially defined by the data set itself and, typically, designed so that the data is dense on the manifold in some sense. The notion of a data space is an abstraction of a manifold encapsulating the essential properties that allow for function approximation. The problem of transfer learning (meta-learning) is to use the learning of a function on one data set to learn a similar function on a new data set. In terms of function approximation, this means lifting a function on one data space (the base data space) to another (the target data space). This viewpoint enables us to connect some inverse problems in applied mathematics (such as inverse Radon transform) with transfer learning. In this paper we examine the question of such lifting when the data is assumed to be known only on a part of the base data space. We are interested in determining subsets of the target data space on which the lifting can be defined, and how the local smoothness of the function and its lifting are related.
☆ TwinExplainer: Explaining Predictions of an Automotive Digital Twin
Vehicles are complex Cyber Physical Systems (CPS) that operate in a variety of environments, and the likelihood of failure of one or more subsystems, such as the engine, transmission, brakes, and fuel, can result in unscheduled downtime and incur high maintenance or repair costs. In order to prevent these issues, it is crucial to continuously monitor the health of various subsystems and identify abnormal sensor channel behavior. Data-driven Digital Twin (DT) systems are capable of such a task. Current DT technologies utilize various Deep Learning (DL) techniques that are constrained by the lack of justification or explanation for their predictions. This inability of these opaque systems can influence decision-making and raises user trust concerns. This paper presents a solution to this issue, where the TwinExplainer system, with its three-layered architectural pipeline, explains the predictions of an automotive DT. Such a system can assist automotive stakeholders in understanding the global scale of the sensor channels and how they contribute towards generic DT predictions. TwinExplainer can also visualize explanations for both normal and abnormal local predictions computed by the DT.
☆ Multi-Grade Deep Learning
The current deep learning model is of a single-grade, that is, it learns a deep neural network by solving a single nonconvex optimization problem. When the layer number of the neural network is large, it is computationally challenging to carry out such a task efficiently. Inspired by the human education process which arranges learning in grades, we propose a multi-grade learning model: We successively solve a number of optimization problems of small sizes, which are organized in grades, to learn a shallow neural network for each grade. Specifically, the current grade is to learn the leftover from the previous grade. In each of the grades, we learn a shallow neural network stacked on the top of the neural network, learned in the previous grades, which remains unchanged in training of the current and future grades. By dividing the task of learning a deep neural network into learning several shallow neural networks, one can alleviate the severity of the nonconvexity of the original optimization problem of a large size. When all grades of the learning are completed, the final neural network learned is a stair-shape neural network, which is the superposition of networks learned from all grades. Such a model enables us to learn a deep neural network much more effectively and efficiently. Moreover, multi-grade learning naturally leads to adaptive learning. We prove that in the context of function approximation if the neural network generated by a new grade is nontrivial, the optimal error of the grade is strictly reduced from the optimal error of the previous grade. Furthermore, we provide several proof-of-concept numerical examples which demonstrate that the proposed multi-grade model outperforms significantly the traditional single-grade model and is much more robust than the traditional model.
♻ ☆ FedBC: Calibrating Global and Local Models via Federated Learning Beyond Consensus
In this work, we quantitatively calibrate the performance of global and local models in federated learning through a multi-criterion optimization-based framework, which we cast as a constrained program. The objective of a device is its local objective, which it seeks to minimize while satisfying nonlinear constraints that quantify the proximity between the local and the global model. By considering the Lagrangian relaxation of this problem, we develop a novel primal-dual method called Federated Learning Beyond Consensus (\texttt{FedBC}). Theoretically, we establish that \texttt{FedBC} converges to a first-order stationary point at rates that matches the state of the art, up to an additional error term that depends on a tolerance parameter introduced to scalarize the multi-criterion formulation. Finally, we demonstrate that \texttt{FedBC} balances the global and local model test accuracy metrics across a suite of datasets (Synthetic, MNIST, CIFAR-10, Shakespeare), achieving competitive performance with state-of-the-art.
♻ ☆ Learning Vortex Dynamics for Fluid Inference and Prediction ICLR 2023
We propose a novel differentiable vortex particle (DVP) method to infer and predict fluid dynamics from a single video. Lying at its core is a particle-based latent space to encapsulate the hidden, Lagrangian vortical evolution underpinning the observable, Eulerian flow phenomena. Our differentiable vortex particles are coupled with a learnable, vortex-to-velocity dynamics mapping to effectively capture the complex flow features in a physically-constrained, low-dimensional space. This representation facilitates the learning of a fluid simulator tailored to the input video that can deliver robust, long-term future predictions. The value of our method is twofold: first, our learned simulator enables the inference of hidden physics quantities (e.g., velocity field) purely from visual observation; secondly, it also supports future prediction, constructing the input video's sequel along with its future dynamics evolution. We compare our method with a range of existing methods on both synthetic and real-world videos, demonstrating improved reconstruction quality, visual plausibility, and physical integrity.
comment: ICLR 2023, project webpage: https://yitongdeng.github.io/vortex_learning_webpage/
♻ ☆ Beyond Exponentially Fast Mixing in Average-Reward Reinforcement Learning via Multi-Level Monte Carlo Actor-Critic
Many existing reinforcement learning (RL) methods employ stochastic gradient iteration on the back end, whose stability hinges upon a hypothesis that the data-generating process mixes exponentially fast with a rate parameter that appears in the step-size selection. Unfortunately, this assumption is violated for large state spaces or settings with sparse rewards, and the mixing time is unknown, making the step size inoperable. In this work, we propose an RL methodology attuned to the mixing time by employing a multi-level Monte Carlo estimator for the critic, the actor, and the average reward embedded within an actor-critic (AC) algorithm. This method, which we call \textbf{M}ulti-level \textbf{A}ctor-\textbf{C}ritic (MAC), is developed especially for infinite-horizon average-reward settings and neither relies on oracle knowledge of the mixing time in its parameter selection nor assumes its exponential decay; it, therefore, is readily applicable to applications with slower mixing times. Nonetheless, it achieves a convergence rate comparable to the state-of-the-art AC algorithms. We experimentally show that these alleviated restrictions on the technical conditions required for stability translate to superior performance in practice for RL problems with sparse rewards.
♻ ☆ Learning Optimal Fair Classification Trees: Trade-offs Between Interpretability, Fairness, and Accuracy
The increasing use of machine learning in high-stakes domains -- where people's livelihoods are impacted -- creates an urgent need for interpretable, fair, and highly accurate algorithms. With these needs in mind, we propose a mixed integer optimization (MIO) framework for learning optimal classification trees -- one of the most interpretable models -- that can be augmented with arbitrary fairness constraints. In order to better quantify the "price of interpretability", we also propose a new measure of model interpretability called decision complexity that allows for comparisons across different classes of machine learning models. We benchmark our method against state-of-the-art approaches for fair classification on popular datasets; in doing so, we conduct one of the first comprehensive analyses of the trade-offs between interpretability, fairness, and predictive accuracy. Given a fixed disparity threshold, our method has a price of interpretability of about 4.2 percentage points in terms of out-of-sample accuracy compared to the best performing, complex models. However, our method consistently finds decisions with almost full parity, while other methods rarely do.
♻ ☆ Posterior Sampling for Continuing Environments
We develop an extension of posterior sampling for reinforcement learning (PSRL) that is suited for a continuing agent-environment interface and integrates naturally into agent designs that scale to complex environments. The approach, continuing PSRL, maintains a statistically plausible model of the environment and follows a policy that maximizes expected $\gamma$-discounted return in that model. At each time, with probability $1-\gamma$, the model is replaced by a sample from the posterior distribution over environments. For a choice of discount factor that suitably depends on the horizon $T$, we establish an $\tilde{O}(\tau S \sqrt{A T})$ bound on the Bayesian regret, where $S$ is the number of environment states, $A$ is the number of actions, and $\tau$ denotes the reward averaging time, which is a bound on the duration required to accurately estimate the average reward of any policy. Our work is the first to formalize and rigorously analyze the resampling approach with randomized exploration.
♻ ☆ Model-agnostic machine learning of conservation laws from data
We present a machine learning based method for learning first integrals of systems of ordinary differential equations from given trajectory data. The method is model-agnostic in that it does not require explicit knowledge of the underlying system of differential equations that generated the trajectories. As a by-product, once the first integrals have been learned, also the system of differential equations will be known. We illustrate our method by considering several classical problems from the mathematical sciences.
comment: 18 pages, 7 figures
♻ ☆ Model-Parallel Fourier Neural Operators as Learned Surrogates for Large-Scale Parametric PDEs
Fourier neural operators (FNOs) are a recently introduced neural network architecture for learning solution operators of partial differential equations (PDEs), which have been shown to perform significantly better than comparable deep learning approaches. Once trained, FNOs can achieve speed-ups of multiple orders of magnitude over conventional numerical PDE solvers. However, due to the high dimensionality of their input data and network weights, FNOs have so far only been applied to two-dimensional or small three-dimensional problems. To remove this limited problem-size barrier, we propose a model-parallel version of FNOs based on domain-decomposition of both the input data and network weights. We demonstrate that our model-parallel FNO is able to predict time-varying PDE solutions of over 2.6 billion variables on Perlmutter using up to 512 A100 GPUs and show an example of training a distributed FNO on the Azure cloud for simulating multiphase CO$_2$ dynamics in the Earth's subsurface.
♻ ☆ Learning to be Fair: A Consequentialist Approach to Equitable Decision-Making
In the dominant paradigm for designing equitable machine learning systems, one works to ensure that model predictions satisfy various fairness criteria, such as parity in error rates across race, gender, and other legally protected traits. That approach, however, typically ignores the downstream decisions and outcomes that predictions affect, and, as a result, can induce unexpected harms. Here we present an alternative framework for fairness that directly anticipates the consequences of decisions. Stakeholders first specify preferences over the possible outcomes of an algorithmically informed decision-making process. For example, lenders may prefer extending credit to those most likely to repay a loan, while also preferring similar lending rates across neighborhoods. One then searches the space of decision policies to maximize the specified utility. We develop and describe a method for efficiently learning these optimal policies from data for a large family of expressive utility functions, facilitating a more holistic approach to equitable decision-making.
♻ ☆ CaSpeR: Latent Spectral Regularization for Continual Learning
While biological intelligence grows organically as new knowledge is gathered throughout life, Artificial Neural Networks forget catastrophically whenever they face a changing training data distribution. Rehearsal-based Continual Learning (CL) approaches have been established as a versatile and reliable solution to overcome this limitation; however, sudden input disruptions and memory constraints are known to alter the consistency of their predictions. We study this phenomenon by investigating the geometric characteristics of the learner's latent space and find that replayed data points of different classes increasingly mix up, interfering with classification. Hence, we propose a geometric regularizer that enforces weak requirements on the Laplacian spectrum of the latent space, promoting a partitioning behavior. We show that our proposal, called Continual Spectral Regularizer (CaSpeR), can be easily combined with any rehearsal-based CL approach and improves the performance of SOTA methods on standard benchmarks. Finally, we conduct additional analysis to provide insights into CaSpeR's effects and applicability.
comment: 11 pages, 3 figures
♻ ☆ Fast and realistic large-scale structure from machine-learning-augmented random field simulations
Producing thousands of simulations of the dark matter distribution in the Universe with increasing precision is a challenging but critical task to facilitate the exploitation of current and forthcoming cosmological surveys. Many inexpensive substitutes to full $N$-body simulations have been proposed, even though they often fail to reproduce the statistics of the smaller, non-linear scales. Among these alternatives, a common approximation is represented by the lognormal distribution, which comes with its own limitations as well, while being extremely fast to compute even for high-resolution density fields. In this work, we train a generative deep learning model, mainly made of convolutional layers, to transform projected lognormal dark matter density fields to more realistic dark matter maps, as obtained from full $N$-body simulations. We detail the procedure that we follow to generate highly correlated pairs of lognormal and simulated maps, which we use as our training data, exploiting the information of the Fourier phases. We demonstrate the performance of our model comparing various statistical tests with different field resolutions, redshifts and cosmological parameters, proving its robustness and explaining its current limitations. When evaluated on 100 test maps, the augmented lognormal random fields reproduce the power spectrum up to wavenumbers of $1 \ h \ \rm{Mpc}^{-1}$, and the bispectrum within 10%, and always within the error bars, of the fiducial target simulations. Finally, we describe how we plan to integrate our proposed model with existing tools to yield more accurate spherical random fields for weak lensing analysis.
comment: 16 pages, 11 figures. Matches MNRAS published version, which includes more tests with e.g. varying cosmological parameters
♻ ☆ Variance, Self-Consistency, and Arbitrariness in Fair Classification
In fair classification, it is common to train a model, and to compare and correct subgroup-specific error rates for disparities. However, even if a model's classification decisions satisfy a fairness metric, it is not necessarily the case that these decisions are equally confident. This becomes clear if we measure variance: We can fix everything in the learning process except the subset of training data, train multiple models, measure (dis)agreement in predictions for each test example, and interpret disagreement to mean that the learning process is more unstable with respect to its classification decision. Empirically, some decisions can in fact be so unstable that they are effectively arbitrary. To reduce this arbitrariness, we formalize a notion of self-consistency of a learning process, develop an ensembling algorithm that provably increases self-consistency, and empirically demonstrate its utility to often improve both fairness and accuracy. Further, our evaluation reveals a startling observation: Applying ensembling to common fair classification benchmarks can significantly reduce subgroup error rate disparities, without employing common pre-, in-, or post-processing fairness interventions. Taken together, our results indicate that variance, particularly on small datasets, can muddle the reliability of conclusions about fairness. One solution is to develop larger benchmark tasks. To this end, we release a toolkit that makes the Home Mortgage Disclosure Act datasets easily usable for future research.
♻ ☆ How Out-of-Distribution Data Hurts Semi-Supervised Learning ICDM'22
Recent semi-supervised learning algorithms have demonstrated greater success with higher overall performance due to better-unlabeled data representations. Nonetheless, recent research suggests that the performance of the SSL algorithm can be degraded when the unlabeled set contains out-of-distribution examples (OODs). This work addresses the following question: How do out-of-distribution (OOD) data adversely affect semi-supervised learning algorithms? To answer this question, we investigate the critical causes of OOD's negative effect on SSL algorithms. In particular, we found that 1) certain kinds of OOD data instances that are close to the decision boundary have a more significant impact on performance than those that are further away, and 2) Batch Normalization (BN), a popular module, may degrade rather than improve performance when the unlabeled set contains OODs. In this context, we developed a unified weighted robust SSL framework that can be easily extended to many existing SSL algorithms and improve their robustness against OODs. More specifically, we developed an efficient bi-level optimization algorithm that could accommodate high-order approximations of the objective and scale to multiple inner optimization steps to learn a massive number of weight parameters while outperforming existing low-order approximations of bi-level optimization. Further, we conduct a theoretical study of the impact of faraway OODs in the BN step and propose a weighted batch normalization (WBN) procedure for improved performance. Finally, we discuss the connection between our approach and low-order approximation techniques. Our experiments on synthetic and real-world datasets demonstrate that our proposed approach significantly enhances the robustness of four representative SSL algorithms against OODs compared to four state-of-the-art robust SSL strategies.
comment: Accepted to ICDM'22
♻ ☆ Sequential Predictive Conformal Inference for Time Series
We present a new distribution-free conformal prediction algorithm for sequential data (e.g., time series), called the \textit{sequential predictive conformal inference} (\texttt{SPCI}). We specifically account for the nature that time series data are non-exchangeable, and thus many existing conformal prediction algorithms are not applicable. The main idea is to exploit the temporal dependence of non-conformity scores (e.g., prediction residuals); thus, the past residuals contain information about future ones. Then we cast the problem of conformal prediction interval as predicting the quantile of a future residual, given a user-specified point prediction algorithm. Theoretically, we establish asymptotic valid conditional coverage upon extending consistency analyses in quantile regression. Using simulation and real-data experiments, we demonstrate a significant reduction in interval width of \texttt{SPCI} compared to other existing methods under the desired empirical coverage.
♻ ☆ Incorporating Sum Constraints into Multitask Gaussian Processes
Machine learning models can be improved by adapting them to respect existing background knowledge. In this paper we consider multitask Gaussian processes, with background knowledge in the form of constraints that require a specific sum of the outputs to be constant. This is achieved by conditioning the prior distribution on the constraint fulfillment. The approach allows for both linear and nonlinear constraints. We demonstrate that the constraints are fulfilled with high precision and that the construction can improve the overall prediction accuracy as compared to the standard Gaussian process.
♻ ☆ Soft Sensing Regression Model: from Sensor to Wafer Metrology Forecasting
The semiconductor industry is one of the most technology-evolving and capital-intensive market sectors. Effective inspection and metrology are necessary to improve product yield, increase product quality and reduce costs. In recent years, many semiconductor manufacturing equipments are equipped with sensors to facilitate real-time monitoring of the production process. These production-state and equipment-state sensor data provide an opportunity to practice machine-learning technologies in various domains, such as anomaly/fault detection, maintenance scheduling, quality prediction, etc. In this work, we focus on the task of soft sensing regression, which uses sensor data to predict impending inspection measurements that used to be measured in wafer inspection and metrology systems. We proposed an LSTM-based regressor and designed two loss functions for model training. Although engineers may look at our prediction errors in a subjective manner, a new piece-wise evaluation metric was proposed for assessing model accuracy in a mathematical way. The experimental results demonstrated that the proposed model can achieve accurate and early prediction of various types of inspections in complicated manufacturing processes.
♻ ☆ Offline Estimation of Controlled Markov Chains: Minimaxity and Sample Complexity
In this work, we study a natural nonparametric estimator of the transition probability matrices of a finite controlled Markov chain. We consider an offline setting with a fixed dataset, collected using a so-called logging policy. We develop sample complexity bounds for the estimator and establish conditions for minimaxity. Our statistical bounds depend on the logging policy through its mixing properties. We show that achieving a particular statistical risk bound involves a subtle and interesting trade-off between the strength of the mixing properties and the number of samples. We demonstrate the validity of our results under various examples, such as ergodic Markov chains, weakly ergodic inhomogeneous Markov chains, and controlled Markov chains with non-stationary Markov, episodic, and greedy controls. Lastly, we use these sample complexity bounds to establish concomitant ones for offline evaluation of stationary Markov control policies.
comment: 67 pages, 23 main-47 appendix
♻ ☆ Pessimistic Off-Policy Optimization for Learning to Rank
Off-policy learning is a framework for optimizing policies without deploying them, using data collected by another policy. In recommender systems, this is especially challenging due to the imbalance in logged data: some items are recommended and thus logged more frequently than others. This is further perpetuated when recommending a list of items, as the action space is combinatorial. To address this challenge, we study pessimistic off-policy optimization for learning to rank. The key idea is to compute lower confidence bounds on parameters of click models and then return the list with the highest pessimistic estimate of its value. This approach is computationally efficient and we analyze it. We study its Bayesian and frequentist variants, and overcome the limitation of unknown prior by incorporating empirical Bayes. To show the empirical effectiveness of our approach, we compare it to off-policy optimizers that use inverse propensity scores or neglect uncertainty. Our approach outperforms all baselines, is robust, and is also general.
♻ ☆ Stream-based active learning with linear models
The proliferation of automated data collection schemes and the advances in sensorics are increasing the amount of data we are able to monitor in real-time. However, given the high annotation costs and the time required by quality inspections, data is often available in an unlabeled form. This is fostering the use of active learning for the development of soft sensors and predictive models. In production, instead of performing random inspections to obtain product information, labels are collected by evaluating the information content of the unlabeled data. Several query strategy frameworks for regression have been proposed in the literature but most of the focus has been dedicated to the static pool-based scenario. In this work, we propose a new strategy for the stream-based scenario, where instances are sequentially offered to the learner, which must instantaneously decide whether to perform the quality check to obtain the label or discard the instance. The approach is inspired by the optimal experimental design theory and the iterative aspect of the decision-making process is tackled by setting a threshold on the informativeness of the unlabeled data points. The proposed approach is evaluated using numerical simulations and the Tennessee Eastman Process simulator. The results confirm that selecting the examples suggested by the proposed algorithm allows for a faster reduction in the prediction error.
comment: Published in Knowledge-Based Systems, final paper available at https://doi.org/10.1016/j.knosys.2022.109664
♻ ☆ Cross-client Label Propagation for Transductive Federated Learning
We present Cross-Client Label Propagation(XCLP), a new method for transductive federated learning. XCLP estimates a data graph jointly from the data of multiple clients and computes labels for the unlabeled data by propagating label information across the graph. To avoid clients having to share their data with anyone, XCLP employs two cryptographically secure protocols: secure Hamming distance computation and secure summation. We demonstrate two distinct applications of XCLP within federated learning. In the first, we use it in a one-shot way to predict labels for unseen test points. In the second, we use it to repeatedly pseudo-label unlabeled training data in a federated semi-supervised setting. Experiments on both real federated and standard benchmark datasets show that in both applications XCLP achieves higher classification accuracy than alternative approaches.
♻ ☆ Complete Neural Networks for Euclidean Graphs
We propose a 2-WL-like geometric graph isomorphism test and prove it is complete when applied to Euclidean Graphs in $\mathbb{R}^3$. We then use recent results on multiset embeddings to devise an efficient geometric GNN model with equivalent separation power. We verify empirically that our GNN model is able to separate particularly challenging synthetic examples, and demonstrate its usefulness for a chemical property prediction problem.
comment: 19 pages, updated 3 figures
♻ ☆ Direct-Effect Risk Minimization for Domain Generalization
We study the problem of out-of-distribution (o.o.d.) generalization where spurious correlations of attributes vary across training and test domains. This is known as the problem of correlation shift and has posed concerns on the reliability of machine learning. In this work, we introduce the concepts of direct and indirect effects from causal inference to the domain generalization problem. We argue that models that learn direct effects minimize the worst-case risk across correlation-shifted domains. To eliminate the indirect effects, our algorithm consists of two stages: in the first stage, we learn an indirect-effect representation by minimizing the prediction error of domain labels using the representation and the class label; in the second stage, we remove the indirect effects learned in the first stage by matching each data with another data of similar indirect-effect representation but of different class label. We also propose a new model selection method by matching the validation set in the same way, which is shown to improve the generalization performance of existing models on correlation-shifted datasets. Experiments on 5 correlation-shifted datasets and the DomainBed benchmark verify the effectiveness of our approach.
♻ ☆ WEASEL 2.0 -- A Random Dilated Dictionary Transform for Fast, Accurate and Memory Constrained Time Series Classification
A time series is a sequence of sequentially ordered real values in time. Time series classification (TSC) is the task of assigning a time series to one of a set of predefined classes, usually based on a model learned from examples. Dictionary-based methods for TSC rely on counting the frequency of certain patterns in time series and are important components of the currently most accurate TSC ensembles. One of the early dictionary-based methods was WEASEL, which at its time achieved SotA results while also being very fast. However, it is outperformed both in terms of speed and accuracy by other methods. Furthermore, its design leads to an unpredictably large memory footprint, making it inapplicable for many applications. In this paper, we present WEASEL 2.0, a complete overhaul of WEASEL based on two recent advancements in TSC: Dilation and ensembling of randomized hyper-parameter settings. These two techniques allow WEASEL 2.0 to work with a fixed-size memory footprint while at the same time improving accuracy. Compared to 15 other SotA methods on the UCR benchmark set, WEASEL 2.0 is significantly more accurate than other dictionary methods and not significantly worse than the currently best methods. Actually, it achieves the highest median accuracy over all data sets, and it performs best in 5 out of 12 problem classes. We thus believe that WEASEL 2.0 is a viable alternative for current TSC and also a potentially interesting input for future ensembles.
♻ ☆ Diffusion-based Image Translation using Disentangled Style and Content Representation ICLR 2023
Diffusion-based image translation guided by semantic texts or a single target image has enabled flexible style transfer which is not limited to the specific domains. Unfortunately, due to the stochastic nature of diffusion models, it is often difficult to maintain the original content of the image during the reverse diffusion. To address this, here we present a novel diffusion-based unsupervised image translation method using disentangled style and content representation. Specifically, inspired by the splicing Vision Transformer, we extract intermediate keys of multihead self attention layer from ViT model and used them as the content preservation loss. Then, an image guided style transfer is performed by matching the [CLS] classification token from the denoised samples and target image, whereas additional CLIP loss is used for the text-driven style transfer. To further accelerate the semantic change during the reverse diffusion, we also propose a novel semantic divergence loss and resampling strategy. Our experimental results show that the proposed method outperforms state-of-the-art baseline models in both text-guided and image-guided translation tasks.
comment: ICLR 2023 camera ready
♻ ☆ A Fast, Well-Founded Approximation to the Empirical Neural Tangent Kernel
Empirical neural tangent kernels (eNTKs) can provide a good understanding of a given network's representation: they are often far less expensive to compute and applicable more broadly than infinite width NTKs. For networks with O output units (e.g. an O-class classifier), however, the eNTK on N inputs is of size $NO \times NO$, taking $O((NO)^2)$ memory and up to $O((NO)^3)$ computation. Most existing applications have therefore used one of a handful of approximations yielding $N \times N$ kernel matrices, saving orders of magnitude of computation, but with limited to no justification. We prove that one such approximation, which we call "sum of logits", converges to the true eNTK at initialization for any network with a wide final "readout" layer. Our experiments demonstrate the quality of this approximation for various uses across a range of settings.
♻ ☆ Max-Quantile Grouped Infinite-Arm Bandits ALT 2023
In this paper, we consider a bandit problem in which there are a number of groups each consisting of infinitely many arms. Whenever a new arm is requested from a given group, its mean reward is drawn from an unknown reservoir distribution (different for each group), and the uncertainty in the arm's mean reward can only be reduced via subsequent pulls of the arm. The goal is to identify the infinite-arm group whose reservoir distribution has the highest $(1-\alpha)$-quantile (e.g., median if $\alpha = \frac{1}{2}$), using as few total arm pulls as possible. We introduce a two-step algorithm that first requests a fixed number of arms from each group and then runs a finite-arm grouped max-quantile bandit algorithm. We characterize both the instance-dependent and worst-case regret, and provide a matching lower bound for the latter, while discussing various strengths, weaknesses, algorithmic improvements, and potential lower bounds associated with our instance-dependent upper bounds.
comment: ALT 2023
♻ ☆ Additive Higher-Order Factorization Machines
In the age of big data and interpretable machine learning, approaches need to work at scale and at the same time allow for a clear mathematical understanding of the method's inner workings. While there exist inherently interpretable semi-parametric regression techniques for large-scale applications to account for non-linearity in the data, their model complexity is still often restricted. One of the main limitations are missing interactions in these models, which are not included for the sake of better interpretability, but also due to untenable computational costs. To address this shortcoming, we derive a scalable high-order tensor product spline model using a factorization approach. Our method allows to include all (higher-order) interactions of non-linear feature effects while having computational costs proportional to a model without interactions. We prove both theoretically and empirically that our methods scales notably better than existing approaches, derive meaningful penalization schemes and also discuss further theoretical aspects. We finally investigate predictive and estimation performance both with synthetic and real data.
♻ ☆ ADAPT : Awesome Domain Adaptation Python Toolbox
In this paper, we introduce the ADAPT library, an open source Python API providing the implementation of the main transfer learning and domain adaptation methods. The library is designed with a user friendly approach to facilitate the access to domain adaptation for a wide public. ADAPT is compatible with scikit-learn and TensorFlow and a full documentation is proposed online https://adapt-python.github.io/adapt/ with a substantial gallery of examples.
comment: 11 pages, 6 figures
♻ ☆ Quantum machine learning beyond kernel methods
Machine learning algorithms based on parametrized quantum circuits are prime candidates for near-term applications on noisy quantum computers. In this direction, various types of quantum machine learning models have been introduced and studied extensively. Yet, our understanding of how these models compare, both mutually and to classical models, remains limited. In this work, we identify a constructive framework that captures all standard models based on parametrized quantum circuits: that of linear quantum models. In particular, we show using tools from quantum information theory how data re-uploading circuits, an apparent outlier of this framework, can be efficiently mapped into the simpler picture of linear models in quantum Hilbert spaces. Furthermore, we analyze the experimentally-relevant resource requirements of these models in terms of qubit number and amount of data needed to learn. Based on recent results from classical machine learning, we prove that linear quantum models must utilize exponentially more qubits than data re-uploading models in order to solve certain learning tasks, while kernel methods additionally require exponentially more data points. Our results provide a more comprehensive view of quantum machine learning models as well as insights on the compatibility of different models with NISQ constraints.
comment: 10+10 pages, 14 figures; significant changes in the main text, corrections in the numerical simulations
♻ ☆ Code2Snapshot: Using Code Snapshots for Learning Representations of Source Code ICML
There are several approaches for encoding source code in the input vectors of neural models. These approaches attempt to include various syntactic and semantic features of input programs in their encoding. In this paper, we investigate Code2Snapshot, a novel representation of the source code that is based on the snapshots of input programs. We evaluate several variations of this representation and compare its performance with state-of-the-art representations that utilize the rich syntactic and semantic features of input programs. Our preliminary study on the utility of Code2Snapshot in the code summarization and code classification tasks suggests that simple snapshots of input programs have comparable performance to state-of-the-art representations. Interestingly, obscuring input programs have insignificant impacts on the Code2Snapshot performance, suggesting that, for some tasks, neural models may provide high performance by relying merely on the structure of input programs.
comment: The 21st IEEE International Conference on Machine Learning and Applications (ICMLA'22)
♻ ☆ FI-ODE: Certified and Robust Forward Invariance in Neural ODEs
Forward invariance is a long-studied property in control theory that is used to certify that a dynamical system stays within some pre-specified set of states for all time, and also admits robustness guarantees (e.g., the certificate holds under perturbations). We propose a general framework for training and provably certifying robust forward invariance in Neural ODEs. We apply this framework in two settings: certified adversarial robustness for image classification, and certified safety in continuous control. Notably, our method empirically produces superior adversarial robustness guarantees compared to prior work on certifiably robust Neural ODEs (including implicit-depth models).
♻ ☆ Decompositional Generation Process for Instance-Dependent Partial Label Learning ICLR 2023
Partial label learning (PLL) is a typical weakly supervised learning problem, where each training example is associated with a set of candidate labels among which only one is true. Most existing PLL approaches assume that the incorrect labels in each training example are randomly picked as the candidate labels and model the generation process of the candidate labels in a simple way. However, these approaches usually do not perform as well as expected due to the fact that the generation process of the candidate labels is always instance-dependent. Therefore, it deserves to be modeled in a refined way. In this paper, we consider instance-dependent PLL and assume that the generation process of the candidate labels could decompose into two sequential parts, where the correct label emerges first in the mind of the annotator but then the incorrect labels related to the feature are also selected with the correct label as candidate labels due to uncertainty of labeling. Motivated by this consideration, we propose a novel PLL method that performs Maximum A Posterior (MAP) based on an explicitly modeled generation process of candidate labels via decomposed probability distribution models. Extensive experiments on manually corrupted benchmark datasets and real-world datasets validate the effectiveness of the proposed method. Source code is available at https://github.com/palm-ml/idgp.
comment: ICLR 2023 Spotlight
♻ ☆ Learning, Fast and Slow: A Goal-Directed Memory-Based Approach for Dynamic Environments
Model-based next state prediction and state value prediction are slow to converge. To address these challenges, we do the following: i) Instead of a neural network, we do model-based planning using a parallel memory retrieval system (which we term the slow mechanism); ii) Instead of learning state values, we guide the agent's actions using goal-directed exploration, by using a neural network to choose the next action given the current state and the goal state (which we term the fast mechanism). The goal-directed exploration is trained online using hippocampal replay of visited states and future imagined states every single time step, leading to fast and efficient training. Empirical studies show that our proposed method has a 92% solve rate across 100 episodes in a dynamically changing grid world, significantly outperforming state-of-the-art actor critic mechanisms such as PPO (54%), TRPO (50%) and A2C (24%). Ablation studies demonstrate that both mechanisms are crucial. We posit that the future of Reinforcement Learning (RL) will be to model goals and sub-goals for various tasks, and plan it out in a goal-directed memory-based approach.
comment: 22 pages
♻ ☆ Towards Answering Open-ended Ethical Quandary Questions
Considerable advancements have been made in various NLP tasks based on the impressive power of large language models (LLMs) and many NLP applications are deployed in our daily lives. In this work, we challenge the capability of LLMs with the new task of Ethical Quandary Generative Question Answering. Ethical quandary questions are more challenging to address because multiple conflicting answers may exist to a single quandary. We explore the current capability of LLMs in providing an answer with a deliberative exchange of different perspectives to an ethical quandary, in the approach of Socratic philosophy, instead of providing a closed answer like an oracle. We propose a model that searches for different ethical principles applicable to the ethical quandary and generates an answer conditioned on the chosen principles through prompt-based few-shot learning. We also discuss the remaining challenges and ethical issues involved in this task and suggest the direction toward developing responsible NLP systems by incorporating human values explicitly.
comment: 16 pages
♻ ☆ Training trajectories, mini-batch losses and the curious role of the learning rate
Stochastic gradient descent plays a fundamental role in nearly all applications of deep learning. However its ability to converge to a global minimum remains shrouded in mystery. In this paper we propose to study the behavior of the loss function on fixed mini-batches along SGD trajectories. We show that the loss function on a fixed batch appears to be remarkably convex-like. In particular for ResNet the loss for any fixed mini-batch can be accurately modeled by a quadratic function and a very low loss value can be reached in just one step of gradient descent with sufficiently large learning rate. We propose a simple model that allows to analyze the relationship between the gradients of stochastic mini-batches and the full batch. Our analysis allows us to discover the equivalency between iterate aggregates and specific learning rate schedules. In particular, for Exponential Moving Average (EMA) and Stochastic Weight Averaging we show that our proposed model matches the observed training trajectories on ImageNet. Our theoretical model predicts that an even simpler averaging technique, averaging just two points a many steps apart, significantly improves accuracy compared to the baseline. We validated our findings on ImageNet and other datasets using ResNet architecture.
comment: 21 pages, 14 figures
♻ ☆ Sliced Optimal Partial Transport
Optimal transport (OT) has become exceedingly popular in machine learning, data science, and computer vision. The core assumption in the OT problem is the equal total amount of mass in source and target measures, which limits its application. Optimal Partial Transport (OPT) is a recently proposed solution to this limitation. Similar to the OT problem, the computation of OPT relies on solving a linear programming problem (often in high dimensions), which can become computationally prohibitive. In this paper, we propose an efficient algorithm for calculating the OPT problem between two non-negative measures in one dimension. Next, following the idea of sliced OT distances, we utilize slicing to define the sliced OPT distance. Finally, we demonstrate the computational and accuracy benefits of the sliced OPT-based method in various numerical experiments. In particular, we show an application of our proposed Sliced-OPT in noisy point cloud registration.
♻ ☆ Distribution free optimality intervals for clustering
We address the problem of validating the ouput of clustering algorithms. Given data $\mathcal{D}$ and a partition $\mathcal{C}$ of these data into $K$ clusters, when can we say that the clusters obtained are correct or meaningful for the data? This paper introduces a paradigm in which a clustering $\mathcal{C}$ is considered meaningful if it is good with respect to a loss function such as the K-means distortion, and stable, i.e. the only good clustering up to small perturbations. Furthermore, we present a generic method to obtain post-inference guarantees of near-optimality and stability for a clustering $\mathcal{C}$. The method can be instantiated for a variety of clustering criteria (also called loss functions) for which convex relaxations exist. Obtaining the guarantees amounts to solving a convex optimization problem. We demonstrate the practical relevance of this method by obtaining guarantees for the K-means and the Normalized Cut clustering criteria on realistic data sets. We also prove that asymptotic instability implies finite sample instability w.h.p., allowing inferences about the population clusterability from a sample. The guarantees do not depend on any distributional assumptions, but they depend on the data set $\mathcal{D}$ admitting a stable clustering.
♻ ☆ ezDPS: An Efficient and Zero-Knowledge Machine Learning Inference Pipeline
Machine Learning as a service (MLaaS) permits resource-limited clients to access powerful data analytics services ubiquitously. Despite its merits, MLaaS poses significant concerns regarding the integrity of delegated computation and the privacy of the server's model parameters. To address this issue, Zhang et al. (CCS'20) initiated the study of zero-knowledge Machine Learning (zkML). Few zkML schemes have been proposed afterward; however, they focus on sole ML classification algorithms that may not offer satisfactory accuracy or require large-scale training data and model parameters, which may not be desirable for some applications. We propose ezDPS, a new efficient and zero-knowledge ML inference scheme. Unlike prior works, ezDPS is a zkML pipeline in which the data is processed in multiple stages for high accuracy. Each stage of ezDPS is harnessed with an established ML algorithm that is shown to be effective in various applications, including Discrete Wavelet Transformation, Principal Components Analysis, and Support Vector Machine. We design new gadgets to prove ML operations effectively. We fully implemented ezDPS and assessed its performance on real datasets. Experimental results showed that ezDPS achieves one-to-three orders of magnitude more efficient than the generic circuit-based approach in all metrics while maintaining more desirable accuracy than single ML classification approaches.
comment: This paper is to appear in Privacy-Enhancing Technologies Symposium (PETS) 2023
♻ ☆ Cognitive Level-$k$ Meta-Learning for Safe and Pedestrian-Aware Autonomous Driving
The potential market for modern self-driving cars is enormous, as they are developing remarkably rapidly. At the same time, however, accidents of pedestrian fatalities caused by autonomous driving have been recorded in the case of street crossing. To ensure traffic safety in self-driving environments and respond to vehicle-human interaction challenges such as jaywalking, we propose Level-$k$ Meta Reinforcement Learning (LK-MRL) algorithm. It takes into account the cognitive hierarchy of pedestrian responses and enables self-driving vehicles to adapt to various human behaviors. %which takes into account pedestrian responses while learning the optimal strategies. As a self-driving vehicle algorithm, the LK-MRL combines level-$k$ thinking into MAML to prepare for heterogeneous pedestrians and improve intersection safety based on the combination of meta-reinforcement learning and human cognitive hierarchy framework. We evaluate the algorithm in two cognitive confrontation hierarchy scenarios in an urban traffic simulator and illustrate its role in ensuring road safety by demonstrating its capability of conjectural and higher-level reasoning.
♻ ☆ Local convexity of the TAP free energy and AMP convergence for Z2-synchronization
We study mean-field variational Bayesian inference using the TAP approach, for Z2-synchronization as a prototypical example of a high-dimensional Bayesian model. We show that for any signal strength $\lambda > 1$ (the weak-recovery threshold), there exists a unique local minimizer of the TAP free energy functional near the mean of the Bayes posterior law. Furthermore, the TAP free energy in a local neighborhood of this minimizer is strongly convex. Consequently, a natural-gradient/mirror-descent algorithm achieves linear convergence to this minimizer from a local initialization, which may be obtained by a constant number of iterates of Approximate Message Passing (AMP). This provides a rigorous foundation for variational inference in high dimensions via minimization of the TAP free energy. We also analyze the finite-sample convergence of AMP, showing that AMP is asymptotically stable at the TAP minimizer for any $\lambda > 1$, and is linearly convergent to this minimizer from a spectral initialization for sufficiently large $\lambda$. Such a guarantee is stronger than results obtainable by state evolution analyses, which only describe a fixed number of AMP iterations in the infinite-sample limit. Our proofs combine the Kac-Rice formula and Sudakov-Fernique Gaussian comparison inequality to analyze the complexity of critical points that satisfy strong convexity and stability conditions within their local neighborhoods.
comment: 62 pages
♻ ☆ The Replicator Dynamic, Chain Components and the Response Graph
In this paper we examine the relationship between the flow of the replicator dynamic, the continuum limit of Multiplicative Weights Update, and a game's response graph. We settle an open problem establishing that under the replicator, sink chain components -- a topological notion of long-run outcome of a dynamical system -- always exist and are approximated by the sink connected components of the game's response graph. More specifically, each sink chain component contains a sink connected component of the response graph, as well as all mixed strategy profiles whose support consists of pure profiles in the same connected component, a set we call the content of the connected component. As a corollary, all profiles are chain recurrent in games with strongly connected response graphs. In any two-player game sharing a response graph with a zero-sum game, the sink chain component is unique. In two-player zero-sum and potential games the sink chain components and sink connected components are in a one-to-one correspondence, and we conjecture that this holds in all games.
comment: 22 pages, 2 figures. Accepted version. To appear in Algorithmic Learning Theory 2023
♻ ☆ Geometry-Complete Perceptron Networks for 3D Molecular Graphs AAAI 2023
The field of geometric deep learning has had a profound impact on the development of innovative and powerful graph neural network architectures. Disciplines such as computer vision and computational biology have benefited significantly from such methodological advances, which has led to breakthroughs in scientific domains such as protein structure prediction and design. In this work, we introduce GCPNet, a new geometry-complete, SE(3)-equivariant graph neural network designed for 3D molecular graph representation learning. We demonstrate the state-of-the-art utility and expressiveness of our method on six independent datasets designed for three distinct geometric tasks: protein-ligand binding affinity prediction, protein structure ranking, and Newtonian many-body systems modeling. Our results suggest that GCPNet is a powerful, general method for capturing complex geometric and physical interactions within 3D molecular graphs for downstream prediction tasks. The source code, data, and instructions to train new models or reproduce our results are freely available at https://github.com/BioinfoMachineLearning/GCPNet.
comment: 18 pages, 3 figures, 11 tables. Under review. Also accepted to DLG-AAAI 2023 and AI2ASE-AAAI 2023. Code available at https://github.com/BioinfoMachineLearning/GCPNet
♻ ☆ Expanding the Deployment Envelope of Behavior Prediction via Adaptive Meta-Learning ICRA 2023
Learning-based behavior prediction methods are increasingly being deployed in real-world autonomous systems, e.g., in fleets of self-driving vehicles, which are beginning to commercially operate in major cities across the world. Despite their advancements, however, the vast majority of prediction systems are specialized to a set of well-explored geographic regions or operational design domains, complicating deployment to additional cities, countries, or continents. Towards this end, we present a novel method for efficiently adapting behavior prediction models to new environments. Our approach leverages recent advances in meta-learning, specifically Bayesian regression, to augment existing behavior prediction models with an adaptive layer that enables efficient domain transfer via offline fine-tuning, online adaptation, or both. Experiments across multiple real-world datasets demonstrate that our method can efficiently adapt to a variety of unseen environments.
comment: 12 pages, 13 figures, 2 tables. To appear at ICRA 2023
♻ ☆ Quality Not Quantity: On the Interaction between Dataset Design and Robustness of CLIP NeurIPS 2022
Web-crawled datasets have enabled remarkable generalization capabilities in recent image-text models such as CLIP (Contrastive Language-Image pre-training) or Flamingo, but little is known about the dataset creation processes. In this work, we introduce a testbed of six publicly available data sources - YFCC, LAION, Conceptual Captions, WIT, RedCaps, Shutterstock - to investigate how pre-training distributions induce robustness in CLIP. We find that the performance of the pre-training data varies substantially across distribution shifts, with no single data source dominating. Moreover, we systematically study the interactions between these data sources and find that combining multiple sources does not necessarily yield better models, but rather dilutes the robustness of the best individual data source. We complement our empirical findings with theoretical insights from a simple setting, where combining the training data also results in diluted robustness. In addition, our theoretical model provides a candidate explanation for the success of the CLIP-based data filtering technique recently employed in the LAION dataset. Overall our results demonstrate that simply gathering a large amount of data from the web is not the most effective way to build a pre-training dataset for robust generalization, necessitating further study into dataset design. Code is available at https://github.com/mlfoundations/clip_quality_not_quantity.
comment: Oral paper at NeurIPS 2022
♻ ☆ CMLCompiler: A Unified Compiler for Classical Machine Learning
Classical machine learning (CML) occupies nearly half of machine learning pipelines in production applications. Unfortunately, it fails to utilize the state-of-the-practice devices fully and performs poorly. Without a unified framework, the hybrid deployments of deep learning (DL) and CML also suffer from severe performance and portability issues. This paper presents the design of a unified compiler, called CMLCompiler, for CML inference. We propose two unified abstractions: operator representations and extended computational graphs. The CMLCompiler framework performs the conversion and graph optimization based on two unified abstractions, then outputs an optimized computational graph to DL compilers or frameworks. We implement CMLCompiler on TVM. The evaluation shows CMLCompiler's portability and superior performance. It achieves up to 4.38x speedup on CPU, 3.31x speedup on GPU, and 5.09x speedup on IoT devices, compared to the state-of-the-art solutions -- scikit-learn, intel sklearn, and hummingbird. Our performance of CML and DL mixed pipelines achieves up to 3.04x speedup compared with cross-framework implementations.
♻ ☆ A Fair Empirical Risk Minimization with Generalized Entropy
This paper studies a parametric family of algorithmic fairness metrics, called generalized entropy, which originally has been used in public welfare and recently introduced to machine learning community. As a meaningful metric to evaluate algorithmic fairness, it requires that generalized entropy specify fairness requirements of a classification problem and the fairness requirements should be realized with small deviation by an algorithm. We investigate the role of generalized entropy as a design parameter for fair classification algorithm through a fair empirical risk minimization with a constraint specified in terms of generalized entropy. We theoretically and experimentally study learnability of the problem.
comment: 56pages and 92 figures Revised for adding experimental results
♻ ☆ Tensor networks for unsupervised machine learning
Modeling the joint distribution of high-dimensional data is a central task in unsupervised machine learning. In recent years, many interests have been attracted to developing learning models based on tensor networks, which have the advantages of a principle understanding of the expressive power using entanglement properties, and as a bridge connecting classical computation and quantum computation. Despite the great potential, however, existing tensor network models for unsupervised machine learning only work as a proof of principle, as their performance is much worse than the standard models such as restricted Boltzmann machines and neural networks. In this Letter, we present autoregressive matrix product states (AMPS), a tensor network model combining matrix product states from quantum many-body physics and autoregressive modeling from machine learning. Our model enjoys the exact calculation of normalized probability and unbiased sampling. We demonstrate the performance of our model using two applications, generative modeling on synthetic and real-world data, and reinforcement learning in statistical physics. Using extensive numerical experiments, we show that the proposed model significantly outperforms the existing tensor network models and the restricted Boltzmann machines, and is competitive with state-of-the-art neural network models.
comment: v2
♻ ☆ Variational Latent Branching Model for Off-Policy Evaluation ICLR 2023
Model-based methods have recently shown great potential for off-policy evaluation (OPE); offline trajectories induced by behavioral policies are fitted to transitions of Markov decision processes (MDPs), which are used to rollout simulated trajectories and estimate the performance of policies. Model-based OPE methods face two key challenges. First, as offline trajectories are usually fixed, they tend to cover limited state and action space. Second, the performance of model-based methods can be sensitive to the initialization of their parameters. In this work, we propose the variational latent branching model (VLBM) to learn the transition function of MDPs by formulating the environmental dynamics as a compact latent space, from which the next states and rewards are then sampled. Specifically, VLBM leverages and extends the variational inference framework with the recurrent state alignment (RSA), which is designed to capture as much information underlying the limited training data, by smoothing out the information flow between the variational (encoding) and generative (decoding) part of VLBM. Moreover, we also introduce the branching architecture to improve the model's robustness against randomly initialized model weights. The effectiveness of the VLBM is evaluated on the deep OPE (DOPE) benchmark, from which the training trajectories are designed to result in varied coverage of the state-action space. We show that the VLBM outperforms existing state-of-the-art OPE methods in general.
comment: Accepted to ICLR 2023
♻ ☆ V2N Service Scaling with Deep Reinforcement Learning
The fifth generation (5G) of wireless networks is set out to meet the stringent requirements of vehicular use cases. Edge computing resources can aid in this direction by moving processing closer to end-users, reducing latency. However, given the stochastic nature of traffic loads and availability of physical resources, appropriate auto-scaling mechanisms need to be employed to support cost-efficient and performant services. To this end, we employ Deep Reinforcement Learning (DRL) for vertical scaling in Edge computing to support vehicular-to-network communications. We address the problem using Deep Deterministic Policy Gradient (DDPG). As DDPG is a model-free off-policy algorithm for learning continuous actions, we introduce a discretization approach to support discrete scaling actions. Thus we address scalability problems inherent to high-dimensional discrete action spaces. Employing a real-world vehicular trace data set, we show that DDPG outperforms existing solutions, reducing (at minimum) the average number of active CPUs by 23% while increasing the long-term reward by 24%.
comment: Accepted at the 36th IEEE/IFIP Network Operations and Management Symposium (NOMS 2023)
♻ ☆ Statistical Inference After Adaptive Sampling for Longitudinal Data
Online reinforcement learning and other adaptive sampling algorithms are increasingly used in digital intervention experiments to optimize treatment delivery for users over time. In this work, we focus on longitudinal user data collected by a large class of adaptive sampling algorithms that are designed to optimize treatment decisions online using accruing data from multiple users. Combining or "pooling" data across users allows adaptive sampling algorithms to potentially learn faster. However, by pooling, these algorithms induce dependence between the collected user data trajectories; we show that this can cause standard variance estimators for i.i.d. data to underestimate the true variance of common estimators on this data type. We develop novel methods to perform a variety of statistical analyses on such adaptively collected data via Z-estimation. Specifically, we introduce the adaptive sandwich variance estimator, a corrected sandwich estimator that leads to consistent variance estimates under adaptive sampling. Additionally, to prove our results we develop significant theory for empirical processes on non-i.i.d., adaptively collected, longitudinal data. This work is motivated by our efforts in designing experiments in which online reinforcement learning algorithms pool data across users to learn to optimize treatment decisions, yet reliable statistical inference is essential for conducting a variety of statistical analyses after the experiment is over.
♻ ☆ Homotopy-based training of NeuralODEs for accurate dynamics discovery ICML2023
Conceptually, Neural Ordinary Differential Equations (NeuralODEs) pose an attractive way to extract dynamical laws from time series data, as they are natural extensions of the traditional differential equation-based modeling paradigm of the physical sciences. In practice, NeuralODEs display long training times and suboptimal results, especially for longer duration data where they may fail to fit the data altogether. While methods have been proposed to stabilize NeuralODE training, many of these involve placing a strong constraint on the functional form the trained NeuralODE can take that the actual underlying governing equation does not guarantee satisfaction. In this work, we present a novel NeuralODE training algorithm that leverages tools from the chaos and mathematical optimization communities - synchronization and homotopy optimization - for a breakthrough in tackling the NeuralODE training obstacle. We demonstrate architectural changes are unnecessary for effective NeuralODE training. Compared to the conventional training methods, our algorithm achieves drastically lower loss values without any changes to the model architectures. Experiments on both simulated and real systems with complex temporal behaviors demonstrate NeuralODEs trained with our algorithm are able to accurately capture true long term behaviors and correctly extrapolate into the future.
comment: 19 pages, 17 figures, submitted to ICML2023
♻ ☆ Private Online Prediction from Experts: Separations and Faster Rates
Online prediction from experts is a fundamental problem in machine learning and several works have studied this problem under privacy constraints. We propose and analyze new algorithms for this problem that improve over the regret bounds of the best existing algorithms for non-adaptive adversaries. For approximate differential privacy, our algorithms achieve regret bounds of $\tilde{O}(\sqrt{T \log d} + \log d/\varepsilon)$ for the stochastic setting and $\tilde O(\sqrt{T \log d} + T^{1/3} \log d/\varepsilon)$ for oblivious adversaries (where $d$ is the number of experts). For pure DP, our algorithms are the first to obtain sub-linear regret for oblivious adversaries in the high-dimensional regime $d \ge T$. Moreover, we prove new lower bounds for adaptive adversaries. Our results imply that unlike the non-private setting, there is a strong separation between the optimal regret for adaptive and non-adaptive adversaries for this problem. Our lower bounds also show a separation between pure and approximate differential privacy for adaptive adversaries where the latter is necessary to achieve the non-private $O(\sqrt{T})$ regret.
comment: Remove the results for the realizable setting which we will upload with additional results for that setting in a separate paper
Multimedia 4
☆ mPLUG-2: A Modularized Multi-modal Foundation Model Across Text, Image and Video
Recent years have witnessed a big convergence of language, vision, and multi-modal pretraining. In this work, we present mPLUG-2, a new unified paradigm with modularized design for multi-modal pretraining, which can benefit from modality collaboration while addressing the problem of modality entanglement. In contrast to predominant paradigms of solely relying on sequence-to-sequence generation or encoder-based instance discrimination, mPLUG-2 introduces a multi-module composition network by sharing common universal modules for modality collaboration and disentangling different modality modules to deal with modality entanglement. It is flexible to select different modules for different understanding and generation tasks across all modalities including text, image, and video. Empirical study shows that mPLUG-2 achieves state-of-the-art or competitive results on a broad range of over 30 downstream tasks, spanning multi-modal tasks of image-text and video-text understanding and generation, and uni-modal tasks of text-only, image-only, and video-only understanding. Notably, mPLUG-2 shows new state-of-the-art results of 48.0 top-1 accuracy and 80.3 CIDEr on the challenging MSRVTT video QA and video caption tasks with a far smaller model size and data scale. It also demonstrates strong zero-shot transferability on vision-language and video-language tasks. Code and models will be released in https://github.com/alibaba/AliceMind.
♻ ☆ Towards Better Quality of Experience in HTTP Adaptive Streaming
HTTP Adaptive Streaming (HAS) is nowadays a popular solution for multimedia delivery. The novelty of HAS lies in the possibility of continuously adapting the streaming session to current network conditions, facilitated by Adaptive Bitrate (ABR) algorithms. Various popular streaming and Video on Demand services such as Netflix, Amazon Prime Video, and Twitch use this method. Given this broad consumer base, ABR algorithms continuously improve to increase user satisfaction. The insights for these improvements are, among others, gathered within the research area of Quality of Experience (QoE). Within this field, various researchers have dedicated their works to identifying potential impairments and testing their impact on viewers' QoE. Two frequently discussed visual impairments influencing QoE are stalling events and quality switches. So far, it is commonly assumed that those stalling events have the worst impact on QoE. This paper challenged this belief and reviewed this assumption by comparing stalling events with multiple quality and high amplitude quality switches. Two subjective studies were conducted. During the first subjective study, participants received a monetary incentive, while the second subjective study was carried out with volunteers. The statistical analysis demonstrated that stalling events do not result in the worst degradation of QoE. These findings suggest that a reevaluation of the effect of stalling events in QoE research is needed. Therefore, these findings may be used for further research and to improve current adaptation strategies in ABR algorithms.
♻ ☆ Image and video compression of fluid flow data
We study the compression of spatial and temporal features in fluid flow data using multimedia compression techniques. The efficacy of spatial compression techniques, including JPEG and JPEG2000 (JP2), and spatio-temporal video compression techniques, namely H.264, H.265, and AV1, in limiting the introduction of compression artifacts and preserving underlying flow physics are considered for laminar periodic wake around a cylinder, two-dimensional turbulence, and turbulent channel flow. These compression techniques significantly compress flow data while maintaining dominant flow features with negligible error. AV1 and H.265 compressions present the best performance across a variety of canonical flow regimes and outperform traditional techniques such as proper orthogonal decomposition in some cases. These image and video compression algorithms are flexible, scalable, and generalizable holding potential for a wide range of applications in fluid dynamics in the context of data storage and transfer.
♻ ☆ Video-Text Modeling with Zero-Shot Transfer from Contrastive Captioners
This work explores an efficient approach to establish a foundational video-text model for tasks including open-vocabulary video classification, text-to-video retrieval, video captioning and video question-answering. We present VideoCoCa that reuses a pretrained image-text contrastive captioner (CoCa) model and adapt it to video-text tasks with minimal extra training. While previous works adapt image-text models with various cross-frame fusion modules (for example, cross-frame attention layer or perceiver resampler) and finetune the modified architecture on video-text data, we surprisingly find that the generative attentional pooling and contrastive attentional pooling layers in the image-text CoCa design are instantly adaptable to "flattened frame embeddings", yielding a strong zero-shot transfer baseline for many video-text tasks. Specifically, the frozen image encoder of a pretrained image-text CoCa takes each video frame as inputs and generates $N$ token embeddings per frame for totally $T$ video frames. We flatten $(N \times T)$ token embeddings as a long sequence of frozen video representation and apply CoCa's generative attentional pooling and contrastive attentional pooling on top. All model weights including pooling layers are directly loaded from an image-text CoCa pretrained model. Without any video or video-text data, VideoCoCa's zero-shot transfer baseline already achieves state-of-the-art results on zero-shot video classification on Kinetics 400/600/700, UCF101, HMDB51, and Charades, as well as zero-shot text-to-video retrieval on MSR-VTT, ActivityNet Captions and VATEX. We also explore lightweight finetuning on top of VideoCoCa, and achieve strong results on video question-answering (iVQA, MSRVTT-QA, MSVD-QA, ActivityNet-QA) and video captioning (MSR-VTT, ActivityNet, VATEX, Youcook2). Our approach establishes a simple video-text baseline for future research.
comment: Tech report. arXiv v2: update results on VATEX, ActivityNet-QA
Computation and Language 49
☆ PADL: Language-Directed Physics-Based Character Control
Developing systems that can synthesize natural and life-like motions for simulated characters has long been a focus for computer animation. But in order for these systems to be useful for downstream applications, they need not only produce high-quality motions, but must also provide an accessible and versatile interface through which users can direct a character's behaviors. Natural language provides a simple-to-use and expressive medium for specifying a user's intent. Recent breakthroughs in natural language processing (NLP) have demonstrated effective use of language-based interfaces for applications such as image generation and program synthesis. In this work, we present PADL, which leverages recent innovations in NLP in order to take steps towards developing language-directed controllers for physics-based character animation. PADL allows users to issue natural language commands for specifying both high-level tasks and low-level skills that a character should perform. We present an adversarial imitation learning approach for training policies to map high-level language commands to low-level controls that enable a character to perform the desired task and skill specified by a user's commands. Furthermore, we propose a multi-task aggregation method that leverages a language-based multiple-choice question-answering approach to determine high-level task objectives from language commands. We show that our framework can be applied to effectively direct a simulated humanoid character to perform a diverse array of complex motor skills.
☆ Mathematical Capabilities of ChatGPT
We investigate the mathematical capabilities of ChatGPT by testing it on publicly available datasets, as well as hand-crafted ones, and measuring its performance against other models trained on a mathematical corpus, such as Minerva. We also test whether ChatGPT can be a useful assistant to professional mathematicians by emulating various use cases that come up in the daily professional activities of mathematicians (question answering, theorem searching). In contrast to formal mathematics, where large databases of formal proofs are available (e.g., the Lean Mathematical Library), current datasets of natural-language mathematics, used to benchmark language models, only cover elementary mathematics. We address this issue by introducing a new dataset: GHOSTS. It is the first natural-language dataset made and curated by working researchers in mathematics that (1) aims to cover graduate-level mathematics and (2) provides a holistic overview of the mathematical capabilities of language models. We benchmark ChatGPT on GHOSTS and evaluate performance against fine-grained criteria. We make this new dataset publicly available to assist a community-driven comparison of ChatGPT with (future) large language models in terms of advanced mathematical comprehension. We conclude that contrary to many positive reports in the media (a potential case of selection bias), ChatGPT's mathematical abilities are significantly below those of an average mathematics graduate student. Our results show that ChatGPT often understands the question but fails to provide correct solutions. Hence, if your goal is to use it to pass a university exam, you would be better off copying from your average peer!
comment: The GHOSTS dataset will be available at https://github.com/friederrr/science-GHOSTS
☆ Benchmarking Large Language Models for News Summarization
Large language models (LLMs) have shown promise for automatic summarization but the reasons behind their successes are poorly understood. By conducting a human evaluation on ten LLMs across different pretraining methods, prompts, and model scales, we make two important observations. First, we find instruction tuning, and not model size, is the key to the LLM's zero-shot summarization capability. Second, existing studies have been limited by low-quality references, leading to underestimates of human performance and lower few-shot and finetuning performance. To better evaluate LLMs, we perform human evaluation over high-quality summaries we collect from freelance writers. Despite major stylistic differences such as the amount of paraphrasing, we find that LMM summaries are judged to be on par with human written summaries.
☆ Do Multi-Document Summarization Models Synthesize? ACL
Multi-document summarization entails producing concise synopses of collections of inputs. For some applications, the synopsis should accurately \emph{synthesize} inputs with respect to a key property or aspect. For example, a synopsis of film reviews all written about a particular movie should reflect the average critic consensus. As a more consequential example, consider narrative summaries that accompany biomedical \emph{systematic reviews} of clinical trial results. These narratives should fairly summarize the potentially conflicting results from individual trials. In this paper we ask: To what extent do modern multi-document summarization models implicitly perform this type of synthesis? To assess this we perform a suite of experiments that probe the degree to which conditional generation models trained for summarization using standard methods yield outputs that appropriately synthesize inputs. We find that existing models do partially perform synthesis, but do so imperfectly. In particular, they are over-sensitive to changes in input ordering and under-sensitive to changes in input compositions (e.g., the ratio of positive to negative movie reviews). We propose a simple, general method for improving model synthesis capabilities by generating an explicitly diverse set of candidate outputs, and then selecting from these the string best aligned with the expected aggregate measure for the inputs, or \emph{abstaining} when the model produces no good candidate. This approach improves model synthesis performance. We hope highlighting the need for synthesis (in some summarization settings), motivates further research into multi-document summarization methods and learning objectives that explicitly account for the need to synthesize.
comment: 22 Pages, 13 Figures, 22 Tables. ACL Formatted paper; expanded version of rejected ICLR submisssion https://openreview.net/forum?id=1PTeB4MWCfU Paper de-anonymized ahead of ICLR de-anonymization due to ACL policies/additional conference submission
☆ Grounding Language Models to Images for Multimodal Generation
We propose an efficient method to ground pretrained text-only language models to the visual domain, enabling them to process and generate arbitrarily interleaved image-and-text data. Our method leverages the abilities of language models learnt from large scale text-only pretraining, such as in-context learning and free-form text generation. We keep the language model frozen, and finetune input and output linear layers to enable cross-modality interactions. This allows our model to process arbitrarily interleaved image-and-text inputs, and generate free-form text interleaved with retrieved images. We achieve strong zero-shot performance on grounded tasks such as contextual image retrieval and multimodal dialogue, and showcase compelling interactive abilities. Our approach works with any off-the-shelf language model and paves the way towards an effective, general solution for leveraging pretrained language models in visually grounded settings.
comment: Project page: https://jykoh.com/fromage
☆ Attend-and-Excite: Attention-Based Semantic Guidance for Text-to-Image Diffusion Models
Recent text-to-image generative models have demonstrated an unparalleled ability to generate diverse and creative imagery guided by a target text prompt. While revolutionary, current state-of-the-art diffusion models may still fail in generating images that fully convey the semantics in the given text prompt. We analyze the publicly available Stable Diffusion model and assess the existence of catastrophic neglect, where the model fails to generate one or more of the subjects from the input prompt. Moreover, we find that in some cases the model also fails to correctly bind attributes (e.g., colors) to their corresponding subjects. To help mitigate these failure cases, we introduce the concept of Generative Semantic Nursing (GSN), where we seek to intervene in the generative process on the fly during inference time to improve the faithfulness of the generated images. Using an attention-based formulation of GSN, dubbed Attend-and-Excite, we guide the model to refine the cross-attention units to attend to all subject tokens in the text prompt and strengthen - or excite - their activations, encouraging the model to generate all subjects described in the text prompt. We compare our approach to alternative approaches and demonstrate that it conveys the desired concepts more faithfully across a range of text prompts.
☆ Execution-based Code Generation using Deep Reinforcement Learning
The utilization of programming language (PL) models, pretrained on large-scale code corpora, as a means of automating software engineering processes has demonstrated considerable potential in streamlining various code generation tasks such as code completion, code translation, and program synthesis. However, current approaches mainly rely on supervised fine-tuning objectives borrowed from text generation, neglecting specific sequence-level features of code, including but not limited to compilability as well as syntactic and functional correctness. To address this limitation, we propose PPOCoder, a new framework for code generation that combines pretrained PL models with Proximal Policy Optimization (PPO) deep reinforcement learning and employs execution feedback as the external source of knowledge into the model optimization. PPOCoder is transferable across different code generation tasks and PLs. Extensive experiments on three code generation tasks demonstrate the effectiveness of our proposed approach compared to SOTA methods, improving the success rate of compilation and functional correctness over different PLs. Our code can be found at https://github.com/reddy-lab-code-research/PPOCoder .
☆ Large Language Models are Versatile Decomposers: Decompose Evidence and Questions for Table-based Reasoning
Table-based reasoning has shown remarkable progress in combining deep models with discrete reasoning, which requires reasoning over both free-form natural language (NL) questions and structured tabular data. However, previous table-based reasoning solutions usually suffer from significant performance degradation on huge evidence (tables). In addition, most existing methods struggle to reason over complex questions since the required information is scattered in different places. To alleviate the above challenges, we exploit large language models (LLMs) as decomposers for effective table-based reasoning, which (i) decompose huge evidence (a huge table) into sub-evidence (a small table) to mitigate the interference of useless information for table reasoning; and (ii) decompose complex questions into simpler sub-questions for text reasoning. Specifically, we first use the LLMs to break down the evidence (tables) involved in the current question, retaining the relevant evidence and excluding the remaining irrelevant evidence from the huge table. In addition, we propose a "parsing-execution-filling" strategy to alleviate the hallucination dilemma of the chain of thought by decoupling logic and numerical computation in each step. Extensive experiments show that our method can effectively leverage decomposed evidence and questions and outperforms the strong baselines on TabFact, WikiTableQuestion, and FetaQA datasets. Notably, our model outperforms human performance for the first time on the TabFact dataset.
☆ The Touché23-ValueEval Dataset for Identifying Human Values behind Arguments
We present the Touch\'e23-ValueEval Dataset for Identifying Human Values behind Arguments. To investigate approaches for the automated detection of human values behind arguments, we collected 9324 arguments from 6 diverse sources, covering religious texts, political discussions, free-text arguments, newspaper editorials, and online democracy platforms. Each argument was annotated by 3 crowdworkers for 54 values. The Touch\'e23-ValueEval dataset extends the Webis-ArgValues-22. In comparison to the previous dataset, the effectiveness of a 1-Baseline decreases, but that of an out-of-the-box BERT model increases. Therefore, though the classification difficulty increased as per the label distribution, the larger dataset allows for training better models.
☆ Dynamic Scheduled Sampling with Imitation Loss for Neural Text Generation
State-of-the-art neural text generation models are typically trained to maximize the likelihood of each token in the ground-truth sequence conditioned on the previous target tokens. However, during inference, the model needs to make a prediction conditioned on the tokens generated by itself. This train-test discrepancy is referred to as exposure bias. Scheduled sampling is a curriculum learning strategy that gradually exposes the model to its own predictions during training to mitigate this bias. Most of the proposed approaches design a scheduler based on training steps, which generally requires careful tuning depending on the training setup. In this work, we introduce Dynamic Scheduled Sampling with Imitation Loss (DySI), which maintains the schedule based solely on the training time accuracy, while enhancing the curriculum learning by introducing an imitation loss, which attempts to make the behavior of the decoder indistinguishable from the behavior of a teacher-forced decoder. DySI is universally applicable across training setups with minimal tuning. Extensive experiments and analysis show that DySI not only achieves notable improvements on standard machine translation benchmarks, but also significantly improves the robustness of other text generation models.
☆ UPop: Unified and Progressive Pruning for Compressing Vision-Language Transformers
Real-world data contains a vast amount of multimodal information, among which vision and language are the two most representative modalities. Moreover, increasingly heavier models, e.g., Transformers, have attracted the attention of researchers to model compression. However, how to compress multimodal models, especially vison-language Transformers, is still under-explored. This paper proposes the \textbf{U}nified and \textbf{P}r\textbf{o}gressive \textbf{P}runing (UPop) as a universal vison-language Transformer compression framework, which incorporates 1) unifiedly searching multimodal subnets in a continuous optimization space from the original model, which enables automatic assignment of pruning ratios among compressible modalities and structures; 2) progressively searching and retraining the subnet, which maintains convergence between the search and retrain to attain higher compression ratios. Experiments on multiple generative and discriminative vision-language tasks, including Visual Reasoning, Image Caption, Visual Question Answer, Image-Text Retrieval, Text-Image Retrieval, and Image Classification, demonstrate the effectiveness and versatility of the proposed UPop framework.
comment: 16 pages, 5 figures, 13 tables
☆ Zero-shot cross-lingual transfer language selection using linguistic similarity
We study the selection of transfer languages for different Natural Language Processing tasks, specifically sentiment analysis, named entity recognition and dependency parsing. In order to select an optimal transfer language, we propose to utilize different linguistic similarity metrics to measure the distance between languages and make the choice of transfer language based on this information instead of relying on intuition. We demonstrate that linguistic similarity correlates with cross-lingual transfer performance for all of the proposed tasks. We also show that there is a statistically significant difference in choosing the optimal language as the transfer source instead of English. This allows us to select a more suitable transfer language which can be used to better leverage knowledge from high-resource languages in order to improve the performance of language applications lacking data. For the study, we used datasets from eight different languages from three language families.
☆ Recursive Neural Networks with Bottlenecks Diagnose (Non-)Compositionality EMNLP 2023
A recent line of work in NLP focuses on the (dis)ability of models to generalise compositionally for artificial languages. However, when considering natural language tasks, the data involved is not strictly, or locally, compositional. Quantifying the compositionality of data is a challenging task, which has been investigated primarily for short utterances. We use recursive neural models (Tree-LSTMs) with bottlenecks that limit the transfer of information between nodes. We illustrate that comparing data's representations in models with and without the bottleneck can be used to produce a compositionality metric. The procedure is applied to the evaluation of arithmetic expressions using synthetic data, and sentiment classification using natural language data. We demonstrate that compression through a bottleneck impacts non-compositional examples disproportionately and then use the bottleneck compositionality metric (BCM) to distinguish compositional from non-compositional samples, yielding a compositionality ranking over a dataset.
comment: Published in EMNLP 2023 findings; 18 pages total (9 in the main paper, 3 pages of limitations and references and 6 pages with appendices)
☆ The Flan Collection: Designing Data and Methods for Effective Instruction Tuning
We study the design decisions of publicly available instruction tuning methods, and break down the development of Flan 2022 (Chung et al., 2022). Through careful ablation studies on the Flan Collection of tasks and methods, we tease apart the effect of design decisions which enable Flan-T5 to outperform prior work by 3-17%+ across evaluation settings. We find task balancing and enrichment techniques are overlooked but critical to effective instruction tuning, and in particular, training with mixed prompt settings (zero-shot, few-shot, and chain-of-thought) actually yields stronger (2%+) performance in all settings. In further experiments, we show Flan-T5 requires less finetuning to converge higher and faster than T5 on single downstream tasks, motivating instruction-tuned models as more computationally-efficient starting checkpoints for new tasks. Finally, to accelerate research on instruction tuning, we make the Flan 2022 collection of datasets, templates, and methods publicly available at https://github.com/google-research/FLAN/tree/main/flan/v2.
☆ Friend-training: Learning from Models of Different but Related Tasks EACL2023
Current self-training methods such as standard self-training, co-training, tri-training, and others often focus on improving model performance on a single task, utilizing differences in input features, model architectures, and training processes. However, many tasks in natural language processing are about different but related aspects of language, and models trained for one task can be great teachers for other related tasks. In this work, we propose friend-training, a cross-task self-training framework, where models trained to do different tasks are used in an iterative training, pseudo-labeling, and retraining process to help each other for better selection of pseudo-labels. With two dialogue understanding tasks, conversational semantic role labeling and dialogue rewriting, chosen for a case study, we show that the models trained with the friend-training framework achieve the best performance compared to strong baselines.
comment: Accepted by EACL2023
☆ Automated Sentiment and Hate Speech Analysis of Facebook Data by Employing Multilingual Transformer Models
In recent years, there has been a heightened consensus within academia and in the public discourse that Social Media Platforms (SMPs), amplify the spread of hateful and negative sentiment content. Researchers have identified how hateful content, political propaganda, and targeted messaging contributed to real-world harms including insurrections against democratically elected governments, genocide, and breakdown of social cohesion due to heightened negative discourse towards certain communities in parts of the world. To counter these issues, SMPs have created semi-automated systems that can help identify toxic speech. In this paper we analyse the statistical distribution of hateful and negative sentiment contents within a representative Facebook dataset (n= 604,703) scrapped through 648 public Facebook pages which identify themselves as proponents (and followers) of far-right Hindutva actors. These pages were identified manually using keyword searches on Facebook and on CrowdTangleand classified as far-right Hindutva pages based on page names, page descriptions, and discourses shared on these pages. We employ state-of-the-art, open-source XLM-T multilingual transformer-based language models to perform sentiment and hate speech analysis of the textual contents shared on these pages over a period of 5.5 years. The result shows the statistical distributions of the predicted sentiment and the hate speech labels; top actors, and top page categories. We further discuss the benchmark performances and limitations of these pre-trained language models.
☆ TopoBERT: Plug and Play Toponym Recognition Module Harnessing Fine-tuned BERT
Extracting precise geographical information from textual contents is crucial in a plethora of applications. For example, during hazardous events, a robust and unbiased toponym extraction framework can provide an avenue to tie the location concerned to the topic discussed by news media posts and pinpoint humanitarian help requests or damage reports from social media. Early studies have leveraged rule-based, gazetteer-based, deep learning, and hybrid approaches to address this problem. However, the performance of existing tools is deficient in supporting operations like emergency rescue, which relies on fine-grained, accurate geographic information. The emerging pretrained language models can better capture the underlying characteristics of text information, including place names, offering a promising pathway to optimize toponym recognition to underpin practical applications. In this paper, TopoBERT, a toponym recognition module based on a one dimensional Convolutional Neural Network (CNN1D) and Bidirectional Encoder Representation from Transformers (BERT), is proposed and fine-tuned. Three datasets (CoNLL2003-Train, Wikipedia3000, WNUT2017) are leveraged to tune the hyperparameters, discover the best training strategy, and train the model. Another two datasets (CoNLL2003-Test and Harvey2017) are used to evaluate the performance. Three distinguished classifiers, linear, multi-layer perceptron, and CNN1D, are benchmarked to determine the optimal model architecture. TopoBERT achieves state-of-the-art performance (f1-score=0.865) compared to the other five baseline models and can be applied to diverse toponym recognition tasks without additional training.
comment: 9 Pages, 6 figures
☆ Archive TimeLine Summarization (ATLS): Conceptual Framework for Timeline Generation over Historical Document Collections
Archive collections are nowadays mostly available through search engines interfaces, which allow a user to retrieve documents by issuing queries. The study of these collections may be, however, impaired by some aspects of search engines, such as the overwhelming number of documents returned or the lack of contextual knowledge provided. New methods that could work independently or in combination with search engines are then required to access these collections. In this position paper, we propose to extend TimeLine Summarization (TLS) methods on archive collections to assist in their studies. We provide an overview of existing TLS methods and we describe a conceptual framework for an Archive TimeLine Summarization (ATLS) system, which aims to generate informative, readable and interpretable timelines.
☆ ZhichunRoad at Amazon KDD Cup 2022: MultiTask Pre-Training for E-Commerce Product Search KDD
In this paper, we propose a robust multilingual model to improve the quality of search results. Our model not only leverage the processed class-balanced dataset, but also benefit from multitask pre-training that leads to more general representations. In pre-training stage, we adopt mlm task, classification task and contrastive learning task to achieve considerably performance. In fine-tuning stage, we use confident learning, exponential moving average method (EMA), adversarial training (FGM) and regularized dropout strategy (R-Drop) to improve the model's generalization and robustness. Moreover, we use a multi-granular semantic unit to discover the queries and products textual metadata for enhancing the representation of the model. Our approach obtained competitive results and ranked top-8 in three tasks. We release the source code and pre-trained models associated with this work.
comment: KDD Cup Workshop @ KDD 2022
☆ Numeracy from Literacy: Data Science as an Emergent Skill from Large Language Models
Large language models (LLM) such as OpenAI's ChatGPT and GPT-3 offer unique testbeds for exploring the translation challenges of turning literacy into numeracy. Previous publicly-available transformer models from eighteen months prior and 1000 times smaller failed to provide basic arithmetic. The statistical analysis of four complex datasets described here combines arithmetic manipulations that cannot be memorized or encoded by simple rules. The work examines whether next-token prediction succeeds from sentence completion into the realm of actual numerical understanding. For example, the work highlights cases for descriptive statistics on in-memory datasets that the LLM initially loads from memory or generates randomly using python libraries. The resulting exploratory data analysis showcases the model's capabilities to group by or pivot categorical sums, infer feature importance, derive correlations, and predict unseen test cases using linear regression. To extend the model's testable range, the research deletes and appends random rows such that recall alone cannot explain emergent numeracy.
☆ Faithful Chain-of-Thought Reasoning
While Chain-of-Thought (CoT) prompting boosts Language Models' (LM) performance on a gamut of complex reasoning tasks, the generated reasoning chain does not necessarily reflect how the model arrives at the answer (aka. faithfulness). We propose Faithful CoT, a faithful-by-construction framework that decomposes a reasoning task into two stages: Translation (Natural Language query $\rightarrow$ symbolic reasoning chain) and Problem Solving (reasoning chain $\rightarrow$ answer), using an LM and a deterministic solver respectively. We demonstrate the efficacy of our approach on 10 reasoning datasets from 4 diverse domains. It outperforms traditional CoT prompting on 9 out of the 10 datasets, with an average accuracy gain of 4.4 on Math Word Problems, 1.9 on Planning, 4.0 on Multi-hop Question Answering (QA), and 18.1 on Logical Inference, under greedy decoding. Together with self-consistency decoding, we achieve new state-of-the-art few-shot performance on 7 out of the 10 datasets, showing a strong synergy between faithfulness and accuracy.
☆ Improving Open-Domain Dialogue Evaluation with a Causal Inference Model
Effective evaluation methods remain a significant challenge for research on open-domain conversational dialogue systems. Explicit satisfaction ratings can be elicited from users, but users often do not provide ratings when asked, and those they give can be highly subjective. Post-hoc ratings by experts are an alternative, but these can be both expensive and complex to collect. Here, we explore the creation of automated methods for predicting both expert and user ratings of open-domain dialogues. We compare four different approaches. First, we train a baseline model using an end-to-end transformer to predict ratings directly from the raw dialogue text. The other three methods are variants of a two-stage approach in which we first extract interpretable features at the turn level that capture, among other aspects, user dialogue behaviors indicating contradiction, repetition, disinterest, compliments, or criticism. We project these features to the dialogue level and train a dialogue-level MLP regression model, a dialogue-level LSTM, and a novel causal inference model called counterfactual-LSTM (CF-LSTM) to predict ratings. The proposed CF-LSTM is a sequential model over turn-level features which predicts ratings using multiple regressors depending on hypotheses derived from the turn-level features. As a causal inference model, CF-LSTM aims to learn the underlying causes of a specific event, such as a low rating. We also bin the user ratings and perform classification experiments with all four models. In evaluation experiments on conversational data from the Alexa Prize SocialBot, we show that the CF-LSTM achieves the best performance for predicting dialogue ratings and classification.
comment: Accepted as a conference paper at IWSDS 2023
☆ Sentence Identification with BOS and EOS Label Combinations EACL 2023
The sentence is a fundamental unit in many NLP applications. Sentence segmentation is widely used as the first preprocessing task, where an input text is split into consecutive sentences considering the end of the sentence (EOS) as their boundaries. This task formulation relies on a strong assumption that the input text consists only of sentences, or what we call the sentential units (SUs). However, real-world texts often contain non-sentential units (NSUs) such as metadata, sentence fragments, nonlinguistic markers, etc. which are unreasonable or undesirable to be treated as a part of an SU. To tackle this issue, we formulate a novel task of sentence identification, where the goal is to identify SUs while excluding NSUs in a given text. To conduct sentence identification, we propose a simple yet effective method which combines the beginning of the sentence (BOS) and EOS labels to determine the most probable SUs and NSUs based on dynamic programming. To evaluate this task, we design an automatic, language-independent procedure to convert the Universal Dependencies corpora into sentence identification benchmarks. Finally, our experiments on the sentence identification task demonstrate that our proposed method generally outperforms sentence segmentation baselines which only utilize EOS labels.
comment: Accepted to EACL 2023 (Findings)
☆ Differentiable Entailment for Parameter Efficient Few Shot Learning
Few-shot learning allows pre-trained language models to adapt to downstream tasks while using a limited number of training examples. However, practical applications are limited when all model parameters must be optimized. In this work we apply a new technique for parameter efficient few shot learning while adopting a strict definition of parameter efficiency. Our training method combines 1) intermediate training by reformulating natural language tasks as entailment tasks \cite{wang_entailment_2021} and 2) differentiable optimization of template and label tokens \cite{zhang_differentiable_2021}. We quantify the tradeoff between parameter efficiency and performance in the few-shot regime and propose a simple model agnostic approach that can be extended to any task By achieving competitive performance while only optimizing 3\% of a model's parameters and allowing for batched inference, we allow for more efficient practical deployment of models.
☆ Universal Topological Regularities of Syntactic Structures: Decoupling Efficiency from Optimization
Human syntactic structures are usually represented as graphs. Much research has focused on the mapping between such graphs and linguistic sequences, but less attention has been paid to the shapes of the graphs themselves: their topologies. This study investigates how the topologies of syntactic graphs reveal traces of the processes that led to their emergence. I report a new universal regularity in syntactic structures: Their topology is communicatively efficient above chance. The pattern holds, without exception, for all 124 languages studied, across linguistic families and modalities (spoken, written, and signed). This pattern can arise from a process optimizing for communicative efficiency or, alternatively, by construction, as a by-effect of a sublinear preferential attachment process reflecting language production mechanisms known from psycholinguistics. This dual explanation shows how communicative efficiency, per se, does not require optimization. Among the two options, efficiency without optimization offers the better explanation for the new pattern.
comment: 30 pages, 7 figures
☆ Machine Translation Impact in E-commerce Multilingual Search EMNLP 2022
Previous work suggests that performance of cross-lingual information retrieval correlates highly with the quality of Machine Translation. However, there may be a threshold beyond which improving query translation quality yields little or no benefit to further improve the retrieval performance. This threshold may depend upon multiple factors including the source and target languages, the existing MT system quality and the search pipeline. In order to identify the benefit of improving an MT system for a given search pipeline, we investigate the sensitivity of retrieval quality to the presence of different levels of MT quality using experimental datasets collected from actual traffic. We systematically improve the performance of our MT systems quality on language pairs as measured by MT evaluation metrics including Bleu and Chrf to determine their impact on search precision metrics and extract signals that help to guide the improvement strategies. Using this information we develop techniques to compare query translations for multiple language pairs and identify the most promising language pairs to invest and improve.
comment: Accepted by EMNLP 2022 (Industry Track)
☆ Detecting Harmful Agendas in News Articles
Manipulated news online is a growing problem which necessitates the use of automated systems to curtail its spread. We argue that while misinformation and disinformation detection have been studied, there has been a lack of investment in the important open challenge of detecting harmful agendas in news articles; identifying harmful agendas is critical to flag news campaigns with the greatest potential for real world harm. Moreover, due to real concerns around censorship, harmful agenda detectors must be interpretable to be effective. In this work, we propose this new task and release a dataset, NewsAgendas, of annotated news articles for agenda identification. We show how interpretable systems can be effective on this task and demonstrate that they can perform comparably to black-box models.
☆ Large Language Models Can Be Easily Distracted by Irrelevant Context
Large language models have achieved impressive performance on various natural language processing tasks. However, so far they have been evaluated primarily on benchmarks where all information in the input context is relevant for solving the task. In this work, we investigate the distractibility of large language models, i.e., how the model problem-solving accuracy can be influenced by irrelevant context. In particular, we introduce Grade-School Math with Irrelevant Context (GSM-IC), an arithmetic reasoning dataset with irrelevant information in the problem description. We use this benchmark to measure the distractibility of cutting-edge prompting techniques for large language models, and find that the model performance is dramatically decreased when irrelevant information is included. We also identify several approaches for mitigating this deficiency, such as decoding with self-consistency and adding to the prompt an instruction that tells the language model to ignore the irrelevant information.
☆ In-Context Retrieval-Augmented Language Models
Retrieval-Augmented Language Modeling (RALM) methods, that condition a language model (LM) on relevant documents from a grounding corpus during generation, have been shown to significantly improve language modeling while also providing a natural source attribution mechanism. Existing RALM approaches focus on modifying the LM architecture in order to facilitate the incorporation of external information, significantly complicating deployment. This paper proposes an under-explored alternative, which we dub In-Context RALM: leaving the LM architecture unchanged and prepending grounding documents to the input. We show that in-context RALM which uses off-the-shelf general purpose retrievers provides surprisingly large LM gains across model sizes and diverse corpora. We also demonstrate that the document retrieval and ranking mechanism can be specialized to the RALM setting to further boost performance. We conclude that in-context RALM has considerable potential to increase the prevalence of LM grounding, particularly in settings where a pretrained LM must be used without modification or even via API access. To that end, we make our code publicly available.
☆ The Power of External Memory in Increasing Predictive Model Capacity
One way of introducing sparsity into deep networks is by attaching an external table of parameters that is sparsely looked up at different layers of the network. By storing the bulk of the parameters in the external table, one can increase the capacity of the model without necessarily increasing the inference time. Two crucial questions in this setting are then: what is the lookup function for accessing the table and how are the contents of the table consumed? Prominent methods for accessing the table include 1) using words/wordpieces token-ids as table indices, 2) LSH hashing the token vector in each layer into a table of buckets, and 3) learnable softmax style routing to a table entry. The ways to consume the contents include adding/concatenating to input representation, and using the contents as expert networks that specialize to different inputs. In this work, we conduct rigorous experimental evaluations of existing ideas and their combinations. We also introduce a new method, alternating updates, that enables access to an increased token dimension without increasing the computation time, and demonstrate its effectiveness in language modeling.
comment: arXiv admin note: substantial text overlap with arXiv:2301.13310
♻ ☆ An Objective Metric for Explainable AI: How and Why to Estimate the Degree of Explainability
Explainable AI was born as a pathway to allow humans to explore and understand the inner working of complex systems. However, establishing what is an explanation and objectively evaluating explainability are not trivial tasks. This paper presents a new model-agnostic metric to measure the Degree of Explainability of information in an objective way. We exploit a specific theoretical model from Ordinary Language Philosophy called the Achinstein's Theory of Explanations, implemented with an algorithm relying on deep language models for knowledge graph extraction and information retrieval. To understand whether this metric can measure explainability, we devised a few experiments and user studies involving more than 190 participants, evaluating two realistic systems for healthcare and finance using famous AI technology, including Artificial Neural Networks and TreeSHAP. The results we obtained are statistically significant (with P values lower than .01), suggesting that our proposed metric for measuring the Degree of Explainability is robust in several scenarios, and it aligns with concrete expectations.
comment: 24 pages, 7 figures, 6 tables, Source code available at: https://github.com/Francesco-Sovrano/DoXpy
♻ ☆ Video Influencers: Unboxing the Mystique
Influencer marketing has become a very popular tool to reach customers. Despite the rapid growth in influencer videos, there has been little research on the effectiveness of their constituent elements in explaining video engagement. We study YouTube influencers and analyze their unstructured video data across text, audio and images using a novel "interpretable deep learning" framework that accomplishes both goals of prediction and interpretation. Our prediction-based approach analyzes unstructured data and finds that "what is said" in words (text) is more influential than "how it is said" in imagery (images) followed by acoustics (audio). Our interpretation-based approach is implemented after completion of model prediction by analyzing the same source of unstructured data to measure importance attributed to the video elements. We eliminate several spurious and confounded relationships, and identify a smaller subset of theory-based relationships. We uncover novel findings that establish distinct effects for measures of shallow and deep engagement which are based on the dual-system framework of human thinking. Our approach is validated using simulated data, and we discuss the learnings from our findings for influencers and brands.
comment: 45 pages, Online Appendix
♻ ☆ Is ChatGPT A Good Translator? A Preliminary Study
This report provides a preliminary evaluation of ChatGPT for machine translation, including translation prompt, multilingual translation, and translation robustness. We adopt the prompts advised by ChatGPT to trigger its translation ability and find that the candidate prompts generally work well and show minor performance differences. By evaluating on a number of benchmark test sets, we find that ChatGPT performs competitively with commercial translation products (e.g., Google Translate) on high-resource European languages but lags behind significantly on low-resource or distant languages. For distant languages, we explore an interesting strategy named $\mathbf{pivot~prompting}$ that asks ChatGPT to translate the source sentence into a high-resource pivot language before into the target language, which improves the translation performance significantly. As for the translation robustness, ChatGPT does not perform as well as the commercial systems on biomedical abstracts or Reddit comments but is potentially a good translator for spoken language. Scripts and data: https://github.com/wxjiao/Is-ChatGPT-A-Good-Translator
comment: 6 pages; added Pivot Prompting for distant languages; added limitations
♻ ☆ Diffuser: Efficient Transformers with Multi-hop Attention Diffusion for Long Sequences
Efficient Transformers have been developed for long sequence modeling, due to their subquadratic memory and time complexity. Sparse Transformer is a popular approach to improving the efficiency of Transformers by restricting self-attention to locations specified by the predefined sparse patterns. However, leveraging sparsity may sacrifice expressiveness compared to full-attention, when important token correlations are multiple hops away. To combine advantages of both the efficiency of sparse transformer and the expressiveness of full-attention Transformer, we propose \textit{Diffuser}, a new state-of-the-art efficient Transformer. Diffuser incorporates all token interactions within one attention layer while maintaining low computation and memory costs. The key idea is to expand the receptive field of sparse attention using Attention Diffusion, which computes multi-hop token correlations based on all paths between corresponding disconnected tokens, besides attention among neighboring tokens. Theoretically, we show the expressiveness of Diffuser as a universal sequence approximator for sequence-to-sequence modeling, and investigate its ability to approximate full-attention by analyzing the graph expander property from the spectral perspective. Experimentally, we investigate the effectiveness of Diffuser with extensive evaluations, including language modeling, image modeling, and Long Range Arena (LRA). Evaluation results show that Diffuser achieves improvements by an average of 0.94% on text classification tasks and 2.30% on LRA, with 1.67$\times$ memory savings compared to state-of-the-art benchmarks, which demonstrates superior performance of Diffuser in both expressiveness and efficiency aspects.
♻ ☆ Automated speech- and text-based classification of neuropsychiatric conditions in a multidiagnostic setting
Speech patterns have been identified as potential diagnostic markers for neuropsychiatric conditions. However, most studies only compare a single clinical group to healthy controls, whereas clinical practice often requires differentiating between multiple potential diagnoses (multiclass settings). To address this, we assembled a dataset of repeated recordings from 420 participants (67 with major depressive disorder, 106 with schizophrenia and 46 with autism, as well as matched controls), and tested the performance of a range of conventional machine learning models and advanced Transformer models on both binary and multiclass classification, based on voice and text features. While binary models performed comparably to previous research (F1 scores between 0.54-0.75 for autism spectrum disorder, ASD; 0.67-0.92 for major depressive disorder, MDD; and 0.71-0.83 for schizophrenia); when differentiating between multiple diagnostic groups performance decreased markedly (F1 scores between 0.35-0.44 for ASD, 0.57-0.75 for MDD, 0.15-0.66 for schizophrenia, and 0.38-0.52 macro F1). Combining voice and text-based models yielded increased performance, suggesting that they capture complementary diagnostic information. Our results indicate that models trained on binary classification may learn to rely on markers of generic differences between clinical and non-clinical populations, or markers of clinical features that overlap across conditions, rather than identifying markers specific to individual conditions. We provide recommendations for future research in the field, suggesting increased focus on developing larger transdiagnostic datasets that include more fine-grained clinical features, and that can support the development of models that better capture the complexity of neuropsychiatric conditions and naturalistic diagnostic assessment.
comment: 24 pages, 5 figures
♻ ☆ The Efficacy of Self-Supervised Speech Models for Audio Representations NeurIPS 2021
Self-supervised learning (SSL) speech models, which can serve as powerful upstream models to extract meaningful speech representations, have achieved unprecedented success in speech representation learning. However, their effectiveness on non-speech datasets is relatively less explored. In this work, we propose an ensemble framework, with a combination of ensemble techniques, to fuse SSL speech models' embeddings. Extensive experiments on speech and non-speech audio datasets are conducted to investigate the representation abilities of our ensemble method and its single constituent model. Ablation studies are carried out to evaluate the performances of different ensemble techniques, such as feature averaging and concatenation. All experiments are conducted during NeurIPS 2021 HEAR Challenge as a standard evaluation pipeline provided by competition officials. Results demonstrate SSL speech models' strong abilities on various non-speech tasks, while we also note that they fail to deal with fine-grained music tasks, such as pitch classification and note onset detection. In addition, feature ensemble is shown to have great potential on producing more holistic representations, as our proposed framework generally surpasses state-of-the-art SSL speech/audio models and has superior performance on various datasets compared with other teams in HEAR Challenge. Our code is available at https://github.com/tony10101105/HEAR-2021-NeurIPS-Challenge -- NTU-GURA.
comment: to appear in Proceedings of Machine Learning Research (PMLR): NeurIPS 2021 Competition Track
♻ ☆ Incorporating Knowledge into Document Summarization: an Application of Prefix-Tuning on GPT-2
Despite the great development of document summarization techniques nowadays, factual inconsistencies between the generated summaries and the original text still occur from time to time. This paper proposes a prefix-tuning-based approach that uses a set of trainable continuous prefix prompt together with discrete prompts to aid model generation, which makes a significant impact on both CNN/Daily Mail and XSum summaries generated using GPT-2. The improvements on fact preservation in the generated summaries indicates the effectiveness of adopting this prefix-tuning-based method in knowledge-enhanced document summarization, and also shows a great potential on other natural language processing tasks.
♻ ☆ One size does not fit all: Investigating strategies for differentially-private learning across NLP tasks EMNLP 2022
Preserving privacy in contemporary NLP models allows us to work with sensitive data, but unfortunately comes at a price. We know that stricter privacy guarantees in differentially-private stochastic gradient descent (DP-SGD) generally degrade model performance. However, previous research on the efficiency of DP-SGD in NLP is inconclusive or even counter-intuitive. In this short paper, we provide an extensive analysis of different privacy preserving strategies on seven downstream datasets in five different `typical' NLP tasks with varying complexity using modern neural models based on BERT and XtremeDistil architectures. We show that unlike standard non-private approaches to solving NLP tasks, where bigger is usually better, privacy-preserving strategies do not exhibit a winning pattern, and each task and privacy regime requires a special treatment to achieve adequate performance.
comment: EMNLP 2022 final camera-ready version
♻ ☆ Difformer: Empowering Diffusion Models on the Embedding Space for Text Generation
Diffusion models have achieved state-of-the-art synthesis quality on both visual and audio tasks, and recent works further adapt them to textual data by diffusing on the embedding space. In this paper, we conduct systematic studies and analyze the challenges between the continuous data space and the embedding space which have not been carefully explored. Firstly, the data distribution is learnable for embeddings, which may lead to the collapse of the loss function. Secondly, as the norm of embeddings varies between popular and rare words, adding the same noise scale will lead to sub-optimal results. In addition, we find the normal level of noise causes insufficient training of the model. To address the above challenges, we propose Difformer, an embedding diffusion model based on Transformer, which consists of three essential modules including an additional anchor loss function, a layer normalization module for embeddings, and a noise factor to the Gaussian noise. Experiments on two seminal text generation tasks including machine translation and text summarization show the superiority of Difformer over compared embedding diffusion baselines.
♻ ☆ Evaluating Unsupervised Text Classification: Zero-shot and Similarity-based Approaches
Text classification of unseen classes is a challenging Natural Language Processing task and is mainly attempted using two different types of approaches. Similarity-based approaches attempt to classify instances based on similarities between text document representations and class description representations. Zero-shot text classification approaches aim to generalize knowledge gained from a training task by assigning appropriate labels of unknown classes to text documents. Although existing studies have already investigated individual approaches to these categories, the experiments in literature do not provide a consistent comparison. This paper addresses this gap by conducting a systematic evaluation of different similarity-based and zero-shot approaches for text classification of unseen classes. Different state-of-the-art approaches are benchmarked on four text classification datasets, including a new dataset from the medical domain. Additionally, novel SimCSE and SBERT-based baselines are proposed, as other baselines used in existing work yield weak classification results and are easily outperformed. Finally, the novel similarity-based Lbl2TransformerVec approach is presented, which outperforms previous state-of-the-art approaches in unsupervised text classification. Our experiments show that similarity-based approaches significantly outperform zero-shot approaches in most cases. Additionally, using SimCSE or SBERT embeddings instead of simpler text representations increases similarity-based classification results even further.
comment: Accepted to 6th International Conference on Natural Language Processing and Information Retrieval (NLPIR '22)
♻ ☆ Towards Better Few-Shot and Finetuning Performance with Forgetful Causal Language Models
Large language models (LLM) trained using the next-token-prediction objective, such as GPT3 and PaLM, have revolutionized natural language processing in recent years by showing impressive zero-shot and few-shot capabilities across a wide range of tasks. In this work, we propose a simple technique that significantly boosts the performance of LLMs without adding computational cost. Our key observation is that, by performing the next token prediction task with randomly selected past tokens masked out, we can improve the quality of the learned representations for downstream language understanding tasks. We hypothesize that randomly masking past tokens prevents over-attending to recent tokens and encourages attention to tokens in the distant past. We find that our method, Forgetful Causal Masking (FCM), significantly improves both few-shot and finetuning performance of PaLM. We further consider a simple extension, T-FCM, which introduces bidirectional context to causal language model without altering the sequence order, and further improves finetuning performance.
comment: Added T-FCM and better FCM results
♻ ☆ Time out of Mind: Generating Rate of Speech conditioned on emotion and speaker
Voice synthesis has seen significant improvements in the past decade resulting in highly intelligible voices. Further investigations have resulted in models that can produce variable speech, including conditional emotional expression. The problem lies, however, in a focus on phrase-level modifications and prosodic vocal features. Using the CREMA-D dataset we have trained a GAN conditioned on emotion to generate worth lengths for a given input text. These word lengths are relative to neutral speech and can be provided, through speech synthesis markup language (SSML) to a text-to-speech (TTS) system to generate more expressive speech. Additionally, a generative model is also trained using implicit maximum likelihood estimation (IMLE) and a comparative analysis with GANs is included. We were able to achieve better performances on objective measures for neutral speech, and better time alignment for happy speech when compared to an out-of-box model. However, further investigation of subjective evaluation is required.
♻ ☆ Bayes risk CTC: Controllable CTC alignment in Sequence-to-Sequence tasks
Sequence-to-Sequence (seq2seq) tasks transcribe the input sequence to a target sequence. The Connectionist Temporal Classification (CTC) criterion is widely used in multiple seq2seq tasks. Besides predicting the target sequence, a side product of CTC is to predict the alignment, which is the most probable input-long sequence that specifies a hard aligning relationship between the input and target units. As there are multiple potential aligning sequences (called paths) that are equally considered in CTC formulation, the choice of which path will be most probable and become the predicted alignment is always uncertain. In addition, it is usually observed that the alignment predicted by vanilla CTC will drift compared with its reference and rarely provides practical functionalities. Thus, the motivation of this work is to make the CTC alignment prediction controllable and thus equip CTC with extra functionalities. The Bayes risk CTC (BRCTC) criterion is then proposed in this work, in which a customizable Bayes risk function is adopted to enforce the desired characteristics of the predicted alignment. With the risk function, the BRCTC is a general framework to adopt some customizable preference over the paths in order to concentrate the posterior into a particular subset of the paths. In applications, we explore one particular preference which yields models with the down-sampling ability and reduced inference costs. By using BRCTC with another preference for early emissions, we obtain an improved performance-latency trade-off for online models. Experimentally, the proposed BRCTC reduces the inference cost of offline models by up to 47% without performance degradation and cuts down the overall latency of online systems to an unseen level.
♻ ☆ SemSup: Semantic Supervision for Simple and Scalable Zero-shot Generalization
Zero-shot learning is the problem of predicting instances over classes not seen during training. One approach to zero-shot learning is providing auxiliary class information to the model. Prior work along this vein have largely used expensive per-instance annotation or singular class-level descriptions, but per-instance descriptions are hard to scale and single class descriptions may not be rich enough. Furthermore, these works have used natural-language descriptions exclusively, simple bi-encoders models, and modality or task-specific methods. These approaches have several limitations: text supervision may not always be available or optimal and bi-encoders may only learn coarse relations between inputs and class descriptions. In this work, we present SemSup, a novel approach that uses (1) a scalable multiple description sampling method which improves performance over single descriptions, (2) alternative description formats such as JSON that are easy to generate and outperform text on certain settings, and (3) hybrid lexical-semantic similarity to leverage fine-grained information in class descriptions. We demonstrate the effectiveness of SemSup across four datasets, two modalities, and three generalization settings. For example, across text and image datasets, SemSup increases unseen class generalization accuracy by 15 points on average compared to the closest baseline.
♻ ☆ PCC: Paraphrasing with Bottom-k Sampling and Cyclic Learning for Curriculum Data Augmentation EACL 2023
Curriculum Data Augmentation (CDA) improves neural models by presenting synthetic data with increasing difficulties from easy to hard. However, traditional CDA simply treats the ratio of word perturbation as the difficulty measure and goes through the curriculums only once. This paper presents \textbf{PCC}: \textbf{P}araphrasing with Bottom-k Sampling and \textbf{C}yclic Learning for \textbf{C}urriculum Data Augmentation, a novel CDA framework via paraphrasing, which exploits the textual paraphrase similarity as the curriculum difficulty measure. We propose a curriculum-aware paraphrase generation module composed of three units: a paraphrase candidate generator with bottom-k sampling, a filtering mechanism and a difficulty measure. We also propose a cyclic learning strategy that passes through the curriculums multiple times. The bottom-k sampling is proposed to generate super-hard instances for the later curriculums. Experimental results on few-shot text classification as well as dialogue generation indicate that PCC surpasses competitive baselines. Human evaluation and extensive case studies indicate that bottom-k sampling effectively generates super-hard instances, and PCC significantly improves the baseline dialogue agent.
comment: Accepted to EACL 2023 (main)
♻ ☆ NLP as a Lens for Causal Analysis and Perception Mining to Infer Mental Health on Social Media
Interactions among humans on social media often convey intentions behind their actions, yielding a psychological language resource for Mental Health Analysis (MHA) of online users. The success of Computational Intelligence Techniques (CIT) for inferring mental illness from such social media resources points to NLP as a lens for causal analysis and perception mining. However, we argue that more consequential and explainable research is required for optimal impact on clinical psychology practice and personalized mental healthcare. To bridge this gap, we posit two significant dimensions: (1) Causal analysis to illustrate a cause and effect relationship in the user generated text; (2) Perception mining to infer psychological perspectives of social effects on online users intentions. Within the scope of Natural Language Processing (NLP), we further explore critical areas of inquiry associated with these two dimensions, specifically through recent advancements in discourse analysis. This position paper guides the community to explore solutions in this space and advance the state of practice in developing conversational agents for inferring mental health from social media. We advocate for a more explainable approach toward modeling computational psychology problems through the lens of language as we observe an increased number of research contributions in dataset and problem formulation for causal relation extraction and perception enhancements while inferring mental states.
comment: Will revise the work
♻ ☆ Towards More Efficient Insertion Transformer with Fractional Positional Encoding
Auto-regressive neural sequence models have been shown to be effective across text generation tasks. However, their left-to-right decoding order prevents generation from being parallelized. Insertion Transformer (Stern et al., 2019) is an attractive alternative that allows outputting multiple tokens in a single generation step. Nevertheless, due to the incompatibility between absolute positional encoding and insertion-based generation schemes, it needs to refresh the encoding of every token in the generated partial hypothesis at each step, which could be costly. We design a novel reusable positional encoding scheme for Insertion Transformers called Fractional Positional Encoding (FPE), which allows reusing representations calculated in previous steps. Empirical studies on various text generation tasks demonstrate the effectiveness of FPE, which leads to floating-point operation reduction and latency improvements on batched decoding.
♻ ☆ WISE: Wavelet Transformation for Boosting Transformers' Long Sequence Learning Ability
Transformer and its variants are fundamental neural architectures in deep learning. Recent works show that learning attention in the Fourier space can improve the long sequence learning capability of Transformers. We argue that wavelet transform shall be a better choice because it captures both position and frequency information with a linear time complexity. Therefore, in this paper, we systematically study the synergy between wavelet transform and Transformers. Specifically, we focus on a new paradigm WISE, which replaces the attention in Transformers by (1) applying forward wavelet transform to project the input sequences to multi-resolution bases, (2) conducting non-linear transformations in the wavelet coefficient space, and (3) reconstructing the representation in input space via backward wavelet transform. Extensive experiments on the Long Range Arena benchmark demonstrate that learning attention in the wavelet space using either fixed or adaptive wavelets can consistently improve Transformer's performance and also significantly outperform Fourier-based methods.
♻ ☆ A Study of Slang Representation Methods
Considering the large amount of content created online by the minute, slang-aware automatic tools are critically needed to promote social good, and assist policymakers and moderators in restricting the spread of offensive language, abuse, and hate speech. Despite the success of large language models and the spontaneous emergence of slang dictionaries, it is unclear how far their combination goes in terms of slang understanding for downstream social good tasks. In this paper, we provide a framework to study different combinations of representation learning models and knowledge resources for a variety of downstream tasks that rely on slang understanding. Our experiments show the superiority of models that have been pre-trained on social media data, while the impact of dictionaries is positive only for static word embeddings. Our error analysis identifies core challenges for slang representation learning, including out-of-vocabulary words, polysemy, variance, and annotation disagreements, which can be traced to characteristics of slang as a quickly evolving and highly subjective language.
Computer Vision and Pattern Recognition 94
☆ From Semi-supervised to Omni-supervised Room Layout Estimation Using Point Clouds ICRA2023
Room layout estimation is a long-existing robotic vision task that benefits both environment sensing and motion planning. However, layout estimation using point clouds (PCs) still suffers from data scarcity due to annotation difficulty. As such, we address the semi-supervised setting of this task based upon the idea of model exponential moving averaging. But adapting this scheme to the state-of-the-art (SOTA) solution for PC-based layout estimation is not straightforward. To this end, we define a quad set matching strategy and several consistency losses based upon metrics tailored for layout quads. Besides, we propose a new online pseudo-label harvesting algorithm that decomposes the distribution of a hybrid distance measure between quads and PC into two components. This technique does not need manual threshold selection and intuitively encourages quads to align with reliable layout points. Surprisingly, this framework also works for the fully-supervised setting, achieving a new SOTA on the ScanNet benchmark. Last but not least, we also push the semi-supervised setting to the realistic omni-supervised setting, demonstrating significantly promoted performance on a newly annotated ARKitScenes testing set. Our codes, data and models are released in this repository.
comment: Accepted to ICRA2023. Code: https://github.com/AIR-DISCOVER/Omni-PQ
☆ Salient Conditional Diffusion for Defending Against Backdoor Attacks
We propose a novel algorithm, Salient Conditional Diffusion (Sancdifi), a state-of-the-art defense against backdoor attacks. Sancdifi uses a denoising diffusion probabilistic model (DDPM) to degrade an image with noise and then recover said image using the learned reverse diffusion. Critically, we compute saliency map-based masks to condition our diffusion, allowing for stronger diffusion on the most salient pixels by the DDPM. As a result, Sancdifi is highly effective at diffusing out triggers in data poisoned by backdoor attacks. At the same time, it reliably recovers salient features when applied to clean data. This performance is achieved without requiring access to the model parameters of the Trojan network, meaning Sancdifi operates as a black-box defense.
comment: 12 pages, 5 figures
☆ Grounding Language Models to Images for Multimodal Generation
We propose an efficient method to ground pretrained text-only language models to the visual domain, enabling them to process and generate arbitrarily interleaved image-and-text data. Our method leverages the abilities of language models learnt from large scale text-only pretraining, such as in-context learning and free-form text generation. We keep the language model frozen, and finetune input and output linear layers to enable cross-modality interactions. This allows our model to process arbitrarily interleaved image-and-text inputs, and generate free-form text interleaved with retrieved images. We achieve strong zero-shot performance on grounded tasks such as contextual image retrieval and multimodal dialogue, and showcase compelling interactive abilities. Our approach works with any off-the-shelf language model and paves the way towards an effective, general solution for leveraging pretrained language models in visually grounded settings.
comment: Project page: https://jykoh.com/fromage
☆ Image Shortcut Squeezing: Countering Perturbative Availability Poisons with Compression
Perturbative availability poisoning (PAP) adds small changes to images to prevent their use for model training. Current research adopts the belief that practical and effective approaches to countering such poisons do not exist. In this paper, we argue that it is time to abandon this belief. We present extensive experiments showing that 12 state-of-the-art PAP methods are vulnerable to Image Shortcut Squeezing (ISS), which is based on simple compression. For example, on average, ISS restores the CIFAR-10 model accuracy to $81.73\%$, surpassing the previous best preprocessing-based countermeasures by $37.97\%$ absolute. ISS also (slightly) outperforms adversarial training and has higher generalizability to unseen perturbation norms and also higher efficiency. Our investigation reveals that the property of PAP perturbations depends on the type of surrogate model used for poison generation, and it explains why a specific ISS compression yields the best performance for a specific type of PAP perturbation. We further test stronger, adaptive poisoning, and show it falls short of being an ideal defense against ISS. Overall, our results demonstrate the importance of considering various (simple) countermeasures to ensure the meaningfulness of analysis carried out during the development of availability poisons.
comment: Our code is available at https://github.com/liuzrcc/ImageShortcutSqueezing
☆ Attend-and-Excite: Attention-Based Semantic Guidance for Text-to-Image Diffusion Models
Recent text-to-image generative models have demonstrated an unparalleled ability to generate diverse and creative imagery guided by a target text prompt. While revolutionary, current state-of-the-art diffusion models may still fail in generating images that fully convey the semantics in the given text prompt. We analyze the publicly available Stable Diffusion model and assess the existence of catastrophic neglect, where the model fails to generate one or more of the subjects from the input prompt. Moreover, we find that in some cases the model also fails to correctly bind attributes (e.g., colors) to their corresponding subjects. To help mitigate these failure cases, we introduce the concept of Generative Semantic Nursing (GSN), where we seek to intervene in the generative process on the fly during inference time to improve the faithfulness of the generated images. Using an attention-based formulation of GSN, dubbed Attend-and-Excite, we guide the model to refine the cross-attention units to attend to all subject tokens in the text prompt and strengthen - or excite - their activations, encouraging the model to generate all subjects described in the text prompt. We compare our approach to alternative approaches and demonstrate that it conveys the desired concepts more faithfully across a range of text prompts.
☆ Patch Gradient Descent: Training Neural Networks on Very Large Images
Traditional CNN models are trained and tested on relatively low resolution images (<300 px), and cannot be directly operated on large-scale images due to compute and memory constraints. We propose Patch Gradient Descent (PatchGD), an effective learning strategy that allows to train the existing CNN architectures on large-scale images in an end-to-end manner. PatchGD is based on the hypothesis that instead of performing gradient-based updates on an entire image at once, it should be possible to achieve a good solution by performing model updates on only small parts of the image at a time, ensuring that the majority of it is covered over the course of iterations. PatchGD thus extensively enjoys better memory and compute efficiency when training models on large scale images. PatchGD is thoroughly evaluated on two datasets - PANDA and UltraMNIST with ResNet50 and MobileNetV2 models under different memory constraints. Our evaluation clearly shows that PatchGD is much more stable and efficient than the standard gradient-descent method in handling large images, and especially when the compute memory is limited.
☆ Ultrasound Based Prosthetic Arm Control
The loss of an upper limb can have a substantial impact on a person's quality of life since it limits a person's ability to work, interact, and perform daily duties independently. Artificial limbs are used in prosthetics to help people who have lost limbs enhance their function and quality of life. Despite significant breakthroughs in prosthetic technology, rejection rates for complex prosthetic devices remain high[1]-[5]. A quarter to a third of upper-limb amputees abandon their prosthetics due to a lack of comprehension of the technology. The most extensively used method for monitoring muscle activity and regulating the prosthetic arm, surface electromyography (sEMG), has significant drawbacks, including a low signal-to-noise ratio and poor amplitude resolution[6]-[8].Unlike myoelectric control systems, which use electrical muscle activation to calculate end-effector velocity, our strategy employs ultrasound to directly monitor mechanical muscle deformation and then uses the extracted signals to proportionally control end-effector location. This investigation made use of four separate hand motions performed by three physically healthy volunteers. A virtual robotic hand simulation was created using ROS. After witnessing performance comparable to that of a hand with very less training, we concluded that our control method is reliable and natural.
☆ Fairness-aware Vision Transformer via Debiased Self-Attention
Vision Transformer (ViT) has recently gained significant interest in solving computer vision (CV) problems due to its capability of extracting informative features and modeling long-range dependencies through the self-attention mechanism. To fully realize the advantages of ViT in real-world applications, recent works have explored the trustworthiness of ViT, including its robustness and explainability. However, another desiderata, fairness has not yet been adequately addressed in the literature. We establish that the existing fairness-aware algorithms (primarily designed for CNNs) do not perform well on ViT. This necessitates the need for developing our novel framework via Debiased Self-Attention (DSA). DSA is a fairness-through-blindness approach that enforces ViT to eliminate spurious features correlated with the sensitive attributes for bias mitigation. Notably, adversarial examples are leveraged to locate and mask the spurious features in the input image patches. In addition, DSA utilizes an attention weights alignment regularizer in the training objective to encourage learning informative features for target prediction. Importantly, our DSA framework leads to improved fairness guarantees over prior works on multiple prediction tasks without compromising target prediction performance
☆ Deep learning-based lung segmentation and automatic regional template in chest X-ray images for pediatric tuberculosis SP
Tuberculosis (TB) is still considered a leading cause of death and a substantial threat to global child health. Both TB infection and disease are curable using antibiotics. However, most children who die of TB are never diagnosed or treated. In clinical practice, experienced physicians assess TB by examining chest X-rays (CXR). Pediatric CXR has specific challenges compared to adult CXR, which makes TB diagnosis in children more difficult. Computer-aided diagnosis systems supported by Artificial Intelligence have shown performance comparable to experienced radiologist TB readings, which could ease mass TB screening and reduce clinical burden. We propose a multi-view deep learning-based solution which, by following a proposed template, aims to automatically regionalize and extract lung and mediastinal regions of interest from pediatric CXR images where key TB findings may be present. Experimental results have shown accurate region extraction, which can be used for further analysis to confirm TB finding presence and severity assessment. Code publicly available at https://github.com/dani-capellan/pTB_LungRegionExtractor.
comment: This work has been accepted at the SPIE Medical Imaging 2023, Image Processing conference
☆ Zero-shot-Learning Cross-Modality Data Translation Through Mutual Information Guided Stochastic Diffusion
Cross-modality data translation has attracted great interest in image computing. Deep generative models (\textit{e.g.}, GANs) show performance improvement in tackling those problems. Nevertheless, as a fundamental challenge in image translation, the problem of Zero-shot-Learning Cross-Modality Data Translation with fidelity remains unanswered. This paper proposes a new unsupervised zero-shot-learning method named Mutual Information guided Diffusion cross-modality data translation Model (MIDiffusion), which learns to translate the unseen source data to the target domain. The MIDiffusion leverages a score-matching-based generative model, which learns the prior knowledge in the target domain. We propose a differentiable local-wise-MI-Layer ($LMI$) for conditioning the iterative denoising sampling. The $LMI$ captures the identical cross-modality features in the statistical domain for the diffusion guidance; thus, our method does not require retraining when the source domain is changed, as it does not rely on any direct mapping between the source and target domains. This advantage is critical for applying cross-modality data translation methods in practice, as a reasonable amount of source domain dataset is not always available for supervised training. We empirically show the advanced performance of MIDiffusion in comparison with an influential group of generative models, including adversarial-based and other score-matching-based models.
☆ UPop: Unified and Progressive Pruning for Compressing Vision-Language Transformers
Real-world data contains a vast amount of multimodal information, among which vision and language are the two most representative modalities. Moreover, increasingly heavier models, e.g., Transformers, have attracted the attention of researchers to model compression. However, how to compress multimodal models, especially vison-language Transformers, is still under-explored. This paper proposes the \textbf{U}nified and \textbf{P}r\textbf{o}gressive \textbf{P}runing (UPop) as a universal vison-language Transformer compression framework, which incorporates 1) unifiedly searching multimodal subnets in a continuous optimization space from the original model, which enables automatic assignment of pruning ratios among compressible modalities and structures; 2) progressively searching and retraining the subnet, which maintains convergence between the search and retrain to attain higher compression ratios. Experiments on multiple generative and discriminative vision-language tasks, including Visual Reasoning, Image Caption, Visual Question Answer, Image-Text Retrieval, Text-Image Retrieval, and Image Classification, demonstrate the effectiveness and versatility of the proposed UPop framework.
comment: 16 pages, 5 figures, 13 tables
☆ A relaxed proximal gradient descent algorithm for convergent plug-and-play with proximal denoiser
This paper presents a new convergent Plug-and-Play (PnP) algorithm. PnP methods are efficient iterative algorithms for solving image inverse problems formulated as the minimization of the sum of a data-fidelity term and a regularization term. PnP methods perform regularization by plugging a pre-trained denoiser in a proximal algorithm, such as Proximal Gradient Descent (PGD). To ensure convergence of PnP schemes, many works study specific parametrizations of deep denoisers. However, existing results require either unverifiable or suboptimal hypotheses on the denoiser, or assume restrictive conditions on the parameters of the inverse problem. Observing that these limitations can be due to the proximal algorithm in use, we study a relaxed version of the PGD algorithm for minimizing the sum of a convex function and a weakly convex one. When plugged with a relaxed proximal denoiser, we show that the proposed PnP-$\alpha$PGD algorithm converges for a wider range of regularization parameters, thus allowing more accurate image restoration.
☆ DisDiff: Unsupervised Disentanglement of Diffusion Probabilistic Models
In this paper, targeting to understand the underlying explainable factors behind observations and modeling the conditional generation process on these factors, we propose a new task, disentanglement of diffusion probabilistic models (DPMs), to take advantage of the remarkable modeling ability of DPMs. To tackle this task, we further devise an unsupervised approach named DisDiff. For the first time, we achieve disentangled representation learning in the framework of diffusion probabilistic models. Given a pre-trained DPM, DisDiff can automatically discover the inherent factors behind the image data and disentangle the gradient fields of DPM into sub-gradient fields, each conditioned on the representation of each discovered factor. We propose a novel Disentangling Loss for DisDiff to facilitate the disentanglement of the representation and sub-gradients. The extensive experiments on synthetic and real-world datasets demonstrate the effectiveness of DisDiff.
☆ Improved distinct bone segmentation in upper-body CT through multi-resolution networks
Purpose: Automated distinct bone segmentation from CT scans is widely used in planning and navigation workflows. U-Net variants are known to provide excellent results in supervised semantic segmentation. However, in distinct bone segmentation from upper body CTs a large field of view and a computationally taxing 3D architecture are required. This leads to low-resolution results lacking detail or localisation errors due to missing spatial context when using high-resolution inputs. Methods: We propose to solve this problem by using end-to-end trainable segmentation networks that combine several 3D U-Nets working at different resolutions. Our approach, which extends and generalizes HookNet and MRN, captures spatial information at a lower resolution and skips the encoded information to the target network, which operates on smaller high-resolution inputs. We evaluated our proposed architecture against single resolution networks and performed an ablation study on information concatenation and the number of context networks. Results: Our proposed best network achieves a median DSC of 0.86 taken over all 125 segmented bone classes and reduces the confusion among similar-looking bones in different locations. These results outperform our previously published 3D U-Net baseline results on the task and distinct-bone segmentation results reported by other groups. Conclusion: The presented multi-resolution 3D U-Nets address current shortcomings in bone segmentation from upper-body CT scans by allowing for capturing a larger field of view while avoiding the cubic growth of the input pixels and intermediate computations that quickly outgrow the computational capacities in 3D. The approach thus improves the accuracy and efficiency of distinct bone segmentation from upper-body CT.
comment: Under submission
☆ What Makes Good Examples for Visual In-Context Learning?
Large-scale models trained on broad data have recently become the mainstream architecture in computer vision due to their strong generalization performance. In this paper, the main focus is on an emergent ability in large vision models, known as in-context learning, which allows inference on unseen tasks by conditioning on in-context examples (a.k.a.~prompt) without updating the model parameters. This concept has been well-known in natural language processing but has only been studied very recently for large vision models. We for the first time provide a comprehensive investigation on the impact of in-context examples in computer vision, and find that the performance is highly sensitive to the choice of in-context examples. To overcome the problem, we propose a prompt retrieval framework to automate the selection of in-context examples. Specifically, we present (1) an unsupervised prompt retrieval method based on nearest example search using an off-the-shelf model, and (2) a supervised prompt retrieval method, which trains a neural network to choose examples that directly maximize in-context learning performance. The results demonstrate that our methods can bring non-trivial improvements to visual in-context learning in comparison to the commonly-used random selection.
comment: code and models:https://github.com/ZhangYuanhan-AI/visual_prompt_retrieval
☆ Spyker: High-performance Library for Spiking Deep Neural Networks
Spiking neural networks (SNNs) have been recently brought to light due to their promising capabilities. SNNs simulate the brain with higher biological plausibility compared to previous generations of neural networks. Learning with fewer samples and consuming less power are among the key features of these networks. However, the theoretical advantages of SNNs have not been seen in practice due to the slowness of simulation tools and the impracticality of the proposed network structures. In this work, we implement a high-performance library named Spyker using C++/CUDA from scratch that outperforms its predecessor. Several SNNs are implemented in this work with different learning rules (spike-timing-dependent plasticity and reinforcement learning) using Spyker that achieve significantly better runtimes, to prove the practicality of the library in the simulation of large-scale networks. To our knowledge, no such tools have been developed to simulate large-scale spiking neural networks with high performance using a modular structure. Furthermore, a comparison of the represented stimuli extracted from Spyker to recorded electrophysiology data is performed to demonstrate the applicability of SNNs in describing the underlying neural mechanisms of the brain functions. The aim of this library is to take a significant step toward uncovering the true potential of the brain computations using SNNs.
comment: 11 pages, 6 figures, 6 listings
☆ A Survey and Benchmark of Automatic Surface Reconstruction from Point Clouds
We survey and benchmark traditional and novel learning-based algorithms that address the problem of surface reconstruction from point clouds. Surface reconstruction from point clouds is particularly challenging when applied to real-world acquisitions, due to noise, outliers, non-uniform sampling and missing data. Traditionally, different handcrafted priors of the input points or the output surface have been proposed to make the problem more tractable. However, hyperparameter tuning for adjusting priors to different acquisition defects can be a tedious task. To this end, the deep learning community has recently addressed the surface reconstruction problem. In contrast to traditional approaches, deep surface reconstruction methods can learn priors directly from a training set of point clouds and corresponding true surfaces. In our survey, we detail how different handcrafted and learned priors affect the robustness of methods to defect-laden input and their capability to generate geometric and topologically accurate reconstructions. In our benchmark, we evaluate the reconstructions of several traditional and learning-based methods on the same grounds. We show that learning-based methods can generalize to unseen shape categories, but their training and test sets must share the same point cloud characteristics. We also provide the code and data to compete in our benchmark and to further stimulate the development of learning-based surface reconstruction https://github.com/raphaelsulzer/dsr-benchmark.
☆ Learning Data Representations with Joint Diffusion Models
We introduce a joint diffusion model that simultaneously learns meaningful internal representations fit for both generative and predictive tasks. Joint machine learning models that allow synthesizing and classifying data often offer uneven performance between those tasks or are unstable to train. In this work, we depart from a set of empirical observations that indicate the usefulness of internal representations built by contemporary deep diffusion-based generative models in both generative and predictive settings. We then introduce an extension of the vanilla diffusion model with a classifier that allows for stable joint training with shared parametrization between those objectives. The resulting joint diffusion model offers superior performance across various tasks, including generative modeling, semi-supervised classification, and domain adaptation.
☆ Priors are Powerful: Improving a Transformer for Multi-camera 3D Detection with 2D Priors
Transfomer-based approaches advance the recent development of multi-camera 3D detection both in academia and industry. In a vanilla transformer architecture, queries are randomly initialised and optimised for the whole dataset, without considering the differences among input frames. In this work, we propose to leverage the predictions from an image backbone, which is often highly optimised for 2D tasks, as priors to the transformer part of a 3D detection network. The method works by (1). augmenting image feature maps with 2D priors, (2). sampling query locations via ray-casting along 2D box centroids, as well as (3). initialising query features with object-level image features. Experimental results shows that 2D priors not only help the model converge faster, but also largely improve the baseline approach by up to 12% in terms of average precision.
☆ Zero3D: Semantic-Driven Multi-Category 3D Shape Generation
Semantic-driven 3D shape generation aims to generate 3D objects conditioned on text. Previous works face problems with single-category generation, low-frequency 3D details, and requiring a large number of paired datasets for training. To tackle these challenges, we propose a multi-category conditional diffusion model. Specifically, 1) to alleviate the problem of lack of large-scale paired data, we bridge the text, 2D image and 3D shape based on the pre-trained CLIP model, and 2) to obtain the multi-category 3D shape feature, we apply the conditional flow model to generate 3D shape vector conditioned on CLIP embedding. 3) to generate multi-category 3D shape, we employ the hidden-layer diffusion model conditioned on the multi-category shape vector, which greatly reduces the training time and memory consumption.
☆ Sport Task: Fine Grained Action Detection and Classification of Table Tennis Strokes from Videos for MediaEval 2022
Sports video analysis is a widespread research topic. Its applications are very diverse, like events detection during a match, video summary, or fine-grained movement analysis of athletes. As part of the MediaEval 2022 benchmarking initiative, this task aims at detecting and classifying subtle movements from sport videos. We focus on recordings of table tennis matches. Conducted since 2019, this task provides a classification challenge from untrimmed videos recorded under natural conditions with known temporal boundaries for each stroke. Since 2021, the task also provides a stroke detection challenge from unannotated, untrimmed videos. This year, the training, validation, and test sets are enhanced to ensure that all strokes are represented in each dataset. The dataset is now similar to the one used in [1, 2]. This research is intended to build tools for coaches and athletes who want to further evaluate their sport performances.
comment: MediaEval 2022 Workshop, Jan 2023, Bergen, Norway. arXiv admin note: substantial text overlap with arXiv:2112.11384
☆ NP-Match: Towards a New Probabilistic Model for Semi-Supervised Learning ICML 2022
Semi-supervised learning (SSL) has been widely explored in recent years, and it is an effective way of leveraging unlabeled data to reduce the reliance on labeled data. In this work, we adjust neural processes (NPs) to the semi-supervised image classification task, resulting in a new method named NP-Match. NP-Match is suited to this task for two reasons. Firstly, NP-Match implicitly compares data points when making predictions, and as a result, the prediction of each unlabeled data point is affected by the labeled data points that are similar to it, which improves the quality of pseudo-labels. Secondly, NP-Match is able to estimate uncertainty that can be used as a tool for selecting unlabeled samples with reliable pseudo-labels. Compared with uncertainty-based SSL methods implemented with Monte-Carlo (MC) dropout, NP-Match estimates uncertainty with much less computational overhead, which can save time at both the training and the testing phases. We conducted extensive experiments on five public datasets under three semi-supervised image classification settings, namely, the standard semi-supervised image classification, the imbalanced semi-supervised image classification, and the multi-label semi-supervised image classification, and NP-Match outperforms state-of-the-art (SOTA) approaches or achieves competitive results on them, which shows the effectiveness of NP-Match and its potential for SSL. The codes are at https://github.com/Jianf-Wang/NP-Match
comment: An journal version of our previous ICML 2022 paper arXiv:2207.01066 . Codes are available at: https://github.com/Jianf-Wang/NP-Match
☆ Lidar Upsampling with Sliced Wasserstein Distance
Lidar became an important component of the perception systems in autonomous driving. But challenges of training data acquisition and annotation made emphasized the role of the sensor to sensor domain adaptation. In this work, we address the problem of lidar upsampling. Learning on lidar point clouds is rather a challenging task due to their irregular and sparse structure. Here we propose a method for lidar point cloud upsampling which can reconstruct fine-grained lidar scan patterns. The key idea is to utilize edge-aware dense convolutions for both feature extraction and feature expansion. Additionally applying a more accurate Sliced Wasserstein Distance facilitates learning of the fine lidar sweep structures. This in turn enables our method to employ a one-stage upsampling paradigm without the need for coarse and fine reconstruction. We conduct several experiments to evaluate our method and demonstrate that it provides better upsampling.
☆ NoiseTransfer: Image Noise Generation with Contrastive Embeddings ACCV 2022
Deep image denoising networks have achieved impressive success with the help of a considerably large number of synthetic train datasets. However, real-world denoising is a still challenging problem due to the dissimilarity between distributions of real and synthetic noisy datasets. Although several real-world noisy datasets have been presented, the number of train datasets (i.e., pairs of clean and real noisy images) is limited, and acquiring more real noise datasets is laborious and expensive. To mitigate this problem, numerous attempts to simulate real noise models using generative models have been studied. Nevertheless, previous works had to train multiple networks to handle multiple different noise distributions. By contrast, we propose a new generative model that can synthesize noisy images with multiple different noise distributions. Specifically, we adopt recent contrastive learning to learn distinguishable latent features of the noise. Moreover, our model can generate new noisy images by transferring the noise characteristics solely from a single reference noisy image. We demonstrate the accuracy and the effectiveness of our noise model for both known and unknown noise removal.
comment: ACCV 2022 oral
Review of methods for automatic cerebral microbleeds detection
Cerebral microbleeds detection is an important and challenging task. With the gaining popularity of the MRI, the ability to detect cerebral microbleeds also raises. Unfortunately, for radiologists, it is a time-consuming and laborious procedure. For this reason, various solutions to automate this process have been proposed for several years, but none of them is currently used in medical practice. In this context, the need to systematize the existing knowledge and best practices has been recognized as a factor facilitating the imminent synthesis of a real CMBs detection system practically applicable in medicine. To the best of our knowledge, all available publications regarding automatic cerebral microbleeds detection have been gathered, described, and assessed in this paper in order to distinguish the current research state and provide a starting point for future studies.
comment: 32 pages, 6 figures, 3 tables, 174 references
☆ AMD: Adaptive Masked Distillation for Object
As a general model compression paradigm, feature-based knowledge distillation allows the student model to learn expressive features from the teacher counterpart. In this paper, we mainly focus on designing an effective feature-distillation framework and propose a spatial-channel adaptive masked distillation (AMD) network for object detection. More specifically, in order to accurately reconstruct important feature regions, we first perform attention-guided feature masking on the feature map of the student network, such that we can identify the important features via spatially adaptive feature masking instead of random masking in the previous methods. In addition, we employ a simple and efficient module to allow the student network channel to be adaptive, improving its model capability in object perception and detection. In contrast to the previous methods, more crucial object-aware features can be reconstructed and learned from the proposed network, which is conducive to accurate object detection. The empirical experiments demonstrate the superiority of our method: with the help of our proposed distillation method, the student networks report 41.3\%, 42.4\%, and 42.7\% mAP scores when RetinaNet, Cascade Mask-RCNN and RepPoints are respectively used as the teacher framework for object detection, which outperforms the previous state-of-the-art distillation methods including FGD and MGD.
☆ Domain-Generalizable Multiple-Domain Clustering
Accurately clustering high-dimensional measurements is vital for adequately analyzing scientific data. Deep learning machinery has remarkably improved clustering capabilities in recent years due to its ability to extract meaningful representations. In this work, we are given unlabeled samples from multiple source domains, and we aim to learn a shared classifier that assigns the examples to various clusters. Evaluation is done by using the classifier for predicting cluster assignments in a previously unseen domain. This setting generalizes the problem of unsupervised domain generalization to the case in which no supervised learning samples are given (completely unsupervised). Towards this goal, we present an end-to-end model and evaluate its capabilities on several multi-domain image datasets. Specifically, we demonstrate that our model is more accurate than schemes that require fine-tuning using samples from the target domain or some level of supervision.
comment: 12 pages, 5 figures
☆ Fourier Sensitivity and Regularization of Computer Vision Models
Recent work has empirically shown that deep neural networks latch on to the Fourier statistics of training data and show increased sensitivity to Fourier-basis directions in the input. Understanding and modifying this Fourier-sensitivity of computer vision models may help improve their robustness. Hence, in this paper we study the frequency sensitivity characteristics of deep neural networks using a principled approach. We first propose a basis trick, proving that unitary transformations of the input-gradient of a function can be used to compute its gradient in the basis induced by the transformation. Using this result, we propose a general measure of any differentiable model's Fourier-sensitivity using the unitary Fourier-transform of its input-gradient. When applied to deep neural networks, we find that computer vision models are consistently sensitive to particular frequencies dependent on the dataset, training method and architecture. Based on this measure, we further propose a Fourier-regularization framework to modify the Fourier-sensitivities and frequency bias of models. Using our proposed regularizer-family, we demonstrate that deep neural networks obtain improved classification accuracy on robustness evaluations.
comment: Published in TMLR, https://openreview.net/forum?id=VmTYgjYloM
☆ Monocular Scene Reconstruction with 3D SDF Transformers ICLR 2023
Monocular scene reconstruction from posed images is challenging due to the complexity of a large environment. Recent volumetric methods learn to directly predict the TSDF volume and have demonstrated promising results in this task. However, most methods focus on how to extract and fuse the 2D features to a 3D feature volume, but none of them improve the way how the 3D volume is aggregated. In this work, we propose an SDF transformer network, which replaces the role of 3D CNN for better 3D feature aggregation. To reduce the explosive computation complexity of the 3D multi-head attention, we propose a sparse window attention module, where the attention is only calculated between the non-empty voxels within a local window. Then a top-down-bottom-up 3D attention network is built for 3D feature aggregation, where a dilate-attention structure is proposed to prevent geometry degeneration, and two global modules are employed to equip with global receptive fields. The experiments on multiple datasets show that this 3D transformer network generates a more accurate and complete reconstruction, which outperforms previous methods by a large margin. Remarkably, the mesh accuracy is improved by 41.8%, and the mesh completeness is improved by 25.3% on the ScanNet dataset. Project page: https://weihaosky.github.io/sdfformer.
comment: Accepted to ICLR 2023
☆ Transfer Learning and Class Decomposition for Detecting the Cognitive Decline of Alzheimer Disease
Early diagnosis of Alzheimer's disease (AD) is essential in preventing the disease's progression. Therefore, detecting AD from neuroimaging data such as structural magnetic resonance imaging (sMRI) has been a topic of intense investigation in recent years. Deep learning has gained considerable attention in Alzheimer's detection. However, training a convolutional neural network from scratch is challenging since it demands more computational time and a significant amount of annotated data. By transferring knowledge learned from other image recognition tasks to medical image classification, transfer learning can provide a promising and effective solution. Irregularities in the dataset distribution present another difficulty. Class decomposition can tackle this issue by simplifying learning a dataset's class boundaries. Motivated by these approaches, this paper proposes a transfer learning method using class decomposition to detect Alzheimer's disease from sMRI images. We use two ImageNet-trained architectures: VGG19 and ResNet50, and an entropy-based technique to determine the most informative images. The proposed model achieved state-of-the-art performance in the Alzheimer's disease (AD) vs mild cognitive impairment (MCI) vs cognitively normal (CN) classification task with a 3\% increase in accuracy from what is reported in the literature.
comment: 12 pages, 3 figures
☆ Adversarial Training of Self-supervised Monocular Depth Estimation against Physical-World Attacks ICLR2023
Monocular Depth Estimation (MDE) is a critical component in applications such as autonomous driving. There are various attacks against MDE networks. These attacks, especially the physical ones, pose a great threat to the security of such systems. Traditional adversarial training method requires ground-truth labels hence cannot be directly applied to self-supervised MDE that does not have ground-truth depth. Some self-supervised model hardening techniques (e.g., contrastive learning) ignore the domain knowledge of MDE and can hardly achieve optimal performance. In this work, we propose a novel adversarial training method for self-supervised MDE models based on view synthesis without using ground-truth depth. We improve adversarial robustness against physical-world attacks using L0-norm-bounded perturbation in training. We compare our method with supervised learning based and contrastive learning based methods that are tailored for MDE. Results on two representative MDE networks show that we achieve better robustness against various adversarial attacks with nearly no benign performance degradation.
comment: Accepted to ICLR2023 (Spotlight)
☆ CRC-RL: A Novel Visual Feature Representation Architecture for Unsupervised Reinforcement Learning
This paper addresses the problem of visual feature representation learning with an aim to improve the performance of end-to-end reinforcement learning (RL) models. Specifically, a novel architecture is proposed that uses a heterogeneous loss function, called CRC loss, to learn improved visual features which can then be used for policy learning in RL. The CRC-loss function is a combination of three individual loss functions, namely, contrastive, reconstruction and consistency loss. The feature representation is learned in parallel to the policy learning while sharing the weight updates through a Siamese Twin encoder model. This encoder model is augmented with a decoder network and a feature projection network to facilitate computation of the above loss components. Through empirical analysis involving latent feature visualization, an attempt is made to provide an insight into the role played by this loss function in learning new action-dependent features and how they are linked to the complexity of the problems being solved. The proposed architecture, called CRC-RL, is shown to outperform the existing state-of-the-art methods on the challenging Deep mind control suite environments by a significant margin thereby creating a new benchmark in this field.
☆ Learning Generalized Hybrid Proximity Representation for Image Recognition ICTAI 2022
Recently, deep metric learning techniques received attention, as the learned distance representations are useful to capture the similarity relationship among samples and further improve the performance of various of supervised or unsupervised learning tasks. We propose a novel supervised metric learning method that can learn the distance metrics in both geometric and probabilistic space for image recognition. In contrast to the previous metric learning methods which usually focus on learning the distance metrics in Euclidean space, our proposed method is able to learn better distance representation in a hybrid approach. To achieve this, we proposed a Generalized Hybrid Metric Loss (GHM-Loss) to learn the general hybrid proximity features from the image data by controlling the trade-off between geometric proximity and probabilistic proximity. To evaluate the effectiveness of our method, we first provide theoretical derivations and proofs of the proposed loss function, then we perform extensive experiments on two public datasets to show the advantage of our method compared to other state-of-the-art metric learning methods.
comment: The paper has been accepted by the IEEE ICTAI 2022
☆ A Survey of Explainable AI in Deep Visual Modeling: Methods and Metrics
Deep visual models have widespread applications in high-stake domains. Hence, their black-box nature is currently attracting a large interest of the research community. We present the first survey in Explainable AI that focuses on the methods and metrics for interpreting deep visual models. Covering the landmark contributions along the state-of-the-art, we not only provide a taxonomic organization of the existing techniques, but also excavate a range of evaluation metrics and collate them as measures of different properties of model explanations. Along the insightful discussion on the current trends, we also discuss the challenges and future avenues for this research direction.
comment: Short accessible survey (9pgs)
☆ Rethinking Soft Label in Label Distribution Learning Perspective
The primary goal of training in early convolutional neural networks (CNN) is the higher generalization performance of the model. However, as the expected calibration error (ECE), which quantifies the explanatory power of model inference, was recently introduced, research on training models that can be explained is in progress. We hypothesized that a gap in supervision criteria during training and inference leads to overconfidence, and investigated that performing label distribution learning (LDL) would enhance the model calibration in CNN training. To verify this assumption, we used a simple LDL setting with recent data augmentation techniques. Based on a series of experiments, the following results are obtained: 1) State-of-the-art KD methods significantly impede model calibration. 2) Training using LDL with recent data augmentation can have excellent effects on model calibration and even in generalization performance. 3) Online LDL brings additional improvements in model calibration and accuracy with long training, especially in large-size models. Using the proposed approach, we simultaneously achieved a lower ECE and higher generalization performance for the image classification datasets CIFAR10, 100, STL10, and ImageNet. We performed several visualizations and analyses and witnessed several interesting behaviors in CNN training with the LDL.
comment: 11 pages main manuscript + references and 11 pages supplementary materials
☆ GeneFace: Generalized and High-Fidelity Audio-Driven 3D Talking Face Synthesis ICLR2023
Generating photo-realistic video portrait with arbitrary speech audio is a crucial problem in film-making and virtual reality. Recently, several works explore the usage of neural radiance field in this task to improve 3D realness and image fidelity. However, the generalizability of previous NeRF-based methods to out-of-domain audio is limited by the small scale of training data. In this work, we propose GeneFace, a generalized and high-fidelity NeRF-based talking face generation method, which can generate natural results corresponding to various out-of-domain audio. Specifically, we learn a variaitional motion generator on a large lip-reading corpus, and introduce a domain adaptative post-net to calibrate the result. Moreover, we learn a NeRF-based renderer conditioned on the predicted facial motion. A head-aware torso-NeRF is proposed to eliminate the head-torso separation problem. Extensive experiments show that our method achieves more generalized and high-fidelity talking face generation compared to previous methods.
comment: Accepted by ICLR2023. Project page: https://geneface.github.io/
☆ Contrast and Clustering: Learning Neighborhood Pair Representation for Source-free Domain Adaptation
Domain adaptation has attracted a great deal of attention in the machine learning community, but it requires access to source data, which often raises concerns about data privacy. We are thus motivated to address these issues and propose a simple yet efficient method. This work treats domain adaptation as an unsupervised clustering problem and trains the target model without access to the source data. Specifically, we propose a loss function called contrast and clustering (CaC), where a positive pair term pulls neighbors belonging to the same class together in the feature space to form clusters, while a negative pair term pushes samples of different classes apart. In addition, extended neighbors are taken into account by querying the nearest neighbor indexes in the memory bank to mine for more valuable negative pairs. Extensive experiments on three common benchmarks, VisDA, Office-Home and Office-31, demonstrate that our method achieves state-of-the-art performance. The code will be made publicly available at https://github.com/yukilulu/CaC.
comment: conference paper
☆ Anomaly Segmentation for High-Resolution Remote Sensing Images Based on Pixel Descriptors AAAI2023
Anomaly segmentation in high spatial resolution (HSR) remote sensing imagery is aimed at segmenting anomaly patterns of the earth deviating from normal patterns, which plays an important role in various Earth vision applications. However, it is a challenging task due to the complex distribution and the irregular shapes of objects, and the lack of abnormal samples. To tackle these problems, an anomaly segmentation model based on pixel descriptors (ASD) is proposed for anomaly segmentation in HSR imagery. Specifically, deep one-class classification is introduced for anomaly segmentation in the feature space with discriminative pixel descriptors. The ASD model incorporates the data argument for generating virtual ab-normal samples, which can force the pixel descriptors to be compact for normal data and meanwhile to be diverse to avoid the model collapse problems when only positive samples participated in the training. In addition, the ASD introduced a multi-level and multi-scale feature extraction strategy for learning the low-level and semantic information to make the pixel descriptors feature-rich. The proposed ASD model was validated using four HSR datasets and compared with the recent state-of-the-art models, showing its potential value in Earth vision applications.
comment: to be published in AAAI2023
☆ Recurrent Structure Attention Guidance for Depth Super-Resolution AAAI-2023
Image guidance is an effective strategy for depth super-resolution. Generally, most existing methods employ hand-crafted operators to decompose the high-frequency (HF) and low-frequency (LF) ingredients from low-resolution depth maps and guide the HF ingredients by directly concatenating them with image features. However, the hand-designed operators usually cause inferior HF maps (e.g., distorted or structurally missing) due to the diverse appearance of complex depth maps. Moreover, the direct concatenation often results in weak guidance because not all image features have a positive effect on the HF maps. In this paper, we develop a recurrent structure attention guided (RSAG) framework, consisting of two important parts. First, we introduce a deep contrastive network with multi-scale filters for adaptive frequency-domain separation, which adopts contrastive networks from large filters to small ones to calculate the pixel contrasts for adaptive high-quality HF predictions. Second, instead of the coarse concatenation guidance, we propose a recurrent structure attention block, which iteratively utilizes the latest depth estimation and the image features to jointly select clear patterns and boundaries, aiming at providing refined guidance for accurate depth recovery. In addition, we fuse the features of HF maps to enhance the edge structures in the decomposed LF maps. Extensive experiments show that our approach obtains superior performance compared with state-of-the-art depth super-resolution methods.
comment: Accepted by AAAI-2023
☆ BRAIxDet: Learning to Detect Malignant Breast Lesion with Incomplete Annotations
Methods to detect malignant lesions from screening mammograms are usually trained with fully annotated datasets, where images are labelled with the localisation and classification of cancerous lesions. However, real-world screening mammogram datasets commonly have a subset that is fully annotated and another subset that is weakly annotated with just the global classification (i.e., without lesion localisation). Given the large size of such datasets, researchers usually face a dilemma with the weakly annotated subset: to not use it or to fully annotate it. The first option will reduce detection accuracy because it does not use the whole dataset, and the second option is too expensive given that the annotation needs to be done by expert radiologists. In this paper, we propose a middle-ground solution for the dilemma, which is to formulate the training as a weakly- and semi-supervised learning problem that we refer to as malignant breast lesion detection with incomplete annotations. To address this problem, our new method comprises two stages, namely: 1) pre-training a multi-view mammogram classifier with weak supervision from the whole dataset, and 2) extending the trained classifier to become a multi-view detector that is trained with semi-supervised student-teacher learning, where the training set contains fully and weakly-annotated mammograms. We provide extensive detection results on two real-world screening mammogram datasets containing incomplete annotations, and show that our proposed approach achieves state-of-the-art results in the detection of malignant breast lesions with incomplete annotations.
☆ Structure Flow-Guided Network for Real Depth Super-Resolution AAAI-2023
Real depth super-resolution (DSR), unlike synthetic settings, is a challenging task due to the structural distortion and the edge noise caused by the natural degradation in real-world low-resolution (LR) depth maps. These defeats result in significant structure inconsistency between the depth map and the RGB guidance, which potentially confuses the RGB-structure guidance and thereby degrades the DSR quality. In this paper, we propose a novel structure flow-guided DSR framework, where a cross-modality flow map is learned to guide the RGB-structure information transferring for precise depth upsampling. Specifically, our framework consists of a cross-modality flow-guided upsampling network (CFUNet) and a flow-enhanced pyramid edge attention network (PEANet). CFUNet contains a trilateral self-attention module combining both the geometric and semantic correlations for reliable cross-modality flow learning. Then, the learned flow maps are combined with the grid-sampling mechanism for coarse high-resolution (HR) depth prediction. PEANet targets at integrating the learned flow map as the edge attention into a pyramid network to hierarchically learn the edge-focused guidance feature for depth edge refinement. Extensive experiments on real and synthetic DSR datasets verify that our approach achieves excellent performance compared to state-of-the-art methods.
comment: Accepted by AAAI-2023
☆ Few-Shot Object Detection via Variational Feature Aggregation AAAI2023
As few-shot object detectors are often trained with abundant base samples and fine-tuned on few-shot novel examples,the learned models are usually biased to base classes and sensitive to the variance of novel examples. To address this issue, we propose a meta-learning framework with two novel feature aggregation schemes. More precisely, we first present a Class-Agnostic Aggregation (CAA) method, where the query and support features can be aggregated regardless of their categories. The interactions between different classes encourage class-agnostic representations and reduce confusion between base and novel classes. Based on the CAA, we then propose a Variational Feature Aggregation (VFA) method, which encodes support examples into class-level support features for robust feature aggregation. We use a variational autoencoder to estimate class distributions and sample variational features from distributions that are more robust to the variance of support examples. Besides, we decouple classification and regression tasks so that VFA is performed on the classification branch without affecting object localization. Extensive experiments on PASCAL VOC and COCO demonstrate that our method significantly outperforms a strong baseline (up to 16\%) and previous state-of-the-art methods (4\% in average). Code will be available at: \url{https://github.com/csuhan/VFA}
comment: Accepted by AAAI2023
☆ A Modular Multi-stage Lightweight Graph Transformer Network for Human Pose and Shape Estimation from 2D Human Pose
In this research, we address the challenge faced by existing deep learning-based human mesh reconstruction methods in balancing accuracy and computational efficiency. These methods typically prioritize accuracy, resulting in large network sizes and excessive computational complexity, which may hinder their practical application in real-world scenarios, such as virtual reality systems. To address this issue, we introduce a modular multi-stage lightweight graph-based transformer network for human pose and shape estimation from 2D human pose, a pose-based human mesh reconstruction approach that prioritizes computational efficiency without sacrificing reconstruction accuracy. Our method consists of a 2D-to-3D lifter module that utilizes graph transformers to analyze structured and implicit joint correlations in 2D human poses, and a mesh regression module that combines the extracted pose features with a mesh template to produce the final human mesh parameters.
☆ ReGANIE: Rectifying GAN Inversion Errors for Accurate Real Image Editing
The StyleGAN family succeed in high-fidelity image generation and allow for flexible and plausible editing of generated images by manipulating the semantic-rich latent style space.However, projecting a real image into its latent space encounters an inherent trade-off between inversion quality and editability. Existing encoder-based or optimization-based StyleGAN inversion methods attempt to mitigate the trade-off but see limited performance. To fundamentally resolve this problem, we propose a novel two-phase framework by designating two separate networks to tackle editing and reconstruction respectively, instead of balancing the two. Specifically, in Phase I, a W-space-oriented StyleGAN inversion network is trained and used to perform image inversion and editing, which assures the editability but sacrifices reconstruction quality. In Phase II, a carefully designed rectifying network is utilized to rectify the inversion errors and perform ideal reconstruction. Experimental results show that our approach yields near-perfect reconstructions without sacrificing the editability, thus allowing accurate manipulation of real images. Further, we evaluate the performance of our rectifying network, and see great generalizability towards unseen manipulation types and out-of-domain images.
☆ Fisheye traffic data set of point center markers
This study presents an open data-market platform and a dataset containing 160,000 markers and 18,000 images. We hope that this dataset will bring more new data value and applications In this paper, we introduce the format and usage of the dataset, and we show a demonstration of deep learning vehicle detection trained by this dataset.
comment: https://youtu.be/sjUQ-Ayxxtk
☆ GaitSADA: Self-Aligned Domain Adaptation for mmWave Gait Recognition
mmWave radar-based gait recognition is a novel user identification method that captures human gait biometrics from mmWave radar return signals. This technology offers privacy protection and is resilient to weather and lighting conditions. However, its generalization performance is yet unknown and limits its practical deployment. To address this problem, in this paper, a non-synthetic dataset is collected and analyzed to reveal the presence of spatial and temporal domain shifts in mmWave gait biometric data, which significantly impacts identification accuracy. To address this issue, a novel self-aligned domain adaptation method called GaitSADA is proposed. GaitSADA improves system generalization performance by using a two-stage semi-supervised model training approach. The first stage uses semi-supervised contrastive learning and the second stage uses semi-supervised consistency training with centroid alignment. Extensive experiments show that GaitSADA outperforms representative domain adaptation methods by an average of 15.41% in low data regimes.
comment: Submitted to ACM Transactions on Sensor Networks (TOSN)
☆ When Source-Free Domain Adaptation Meets Learning with Noisy Labels ICLR 2023
Recent state-of-the-art source-free domain adaptation (SFDA) methods have focused on learning meaningful cluster structures in the feature space, which have succeeded in adapting the knowledge from source domain to unlabeled target domain without accessing the private source data. However, existing methods rely on the pseudo-labels generated by source models that can be noisy due to domain shift. In this paper, we study SFDA from the perspective of learning with label noise (LLN). Unlike the label noise in the conventional LLN scenario, we prove that the label noise in SFDA follows a different distribution assumption. We also prove that such a difference makes existing LLN methods that rely on their distribution assumptions unable to address the label noise in SFDA. Empirical evidence suggests that only marginal improvements are achieved when applying the existing LLN methods to solve the SFDA problem. On the other hand, although there exists a fundamental difference between the label noise in the two scenarios, we demonstrate theoretically that the early-time training phenomenon (ETP), which has been previously observed in conventional label noise settings, can also be observed in the SFDA problem. Extensive experiments demonstrate significant improvements to existing SFDA algorithms by leveraging ETP to address the label noise in SFDA.
comment: 33 pages, 16 figures, accepted by ICLR 2023
☆ Quantized Neural Networks for Low-Precision Accumulation with Guaranteed Overflow Avoidance
We introduce a quantization-aware training algorithm that guarantees avoiding numerical overflow when reducing the precision of accumulators during inference. We leverage weight normalization as a means of constraining parameters during training using accumulator bit width bounds that we derive. We evaluate our algorithm across multiple quantized models that we train for different tasks, showing that our approach can reduce the precision of accumulators while maintaining model accuracy with respect to a floating-point baseline. We then show that this reduction translates to increased design efficiency for custom FPGA-based accelerators. Finally, we show that our algorithm not only constrains weights to fit into an accumulator of user-defined bit width, but also increases the sparsity and compressibility of the resulting weights. Across all of our benchmark models trained with 8-bit weights and activations, we observe that constraining the hidden layers of quantized neural networks to fit into 16-bit accumulators yields an average 98.2% sparsity with an estimated compression rate of 46.5x all while maintaining 99.2% of the floating-point performance.
☆ Demystifying Disagreement-on-the-Line in High Dimensions
Evaluating the performance of machine learning models under distribution shift is challenging, especially when we only have unlabeled data from the shifted (target) domain, along with labeled data from the original (source) domain. Recent work suggests that the notion of disagreement, the degree to which two models trained with different randomness differ on the same input, is a key to tackle this problem. Experimentally, disagreement and prediction error have been shown to be strongly connected, which has been used to estimate model performance. Experiments have lead to the discovery of the disagreement-on-the-line phenomenon, whereby the classification error under the target domain is often a linear function of the classification error under the source domain; and whenever this property holds, disagreement under the source and target domain follow the same linear relation. In this work, we develop a theoretical foundation for analyzing disagreement in high-dimensional random features regression; and study under what conditions the disagreement-on-the-line phenomenon occurs in our setting. Experiments on CIFAR-10-C, Tiny ImageNet-C, and Camelyon17 are consistent with our theory and support the universality of the theoretical findings.
☆ CaraNet: Context Axial Reverse Attention Network for Segmentation of Small Medical Objects
Segmenting medical images accurately and reliably is important for disease diagnosis and treatment. It is a challenging task because of the wide variety of objects' sizes, shapes, and scanning modalities. Recently, many convolutional neural networks (CNN) have been designed for segmentation tasks and achieved great success. Few studies, however, have fully considered the sizes of objects, and thus most demonstrate poor performance for small objects segmentation. This can have a significant impact on the early detection of diseases. This paper proposes a Context Axial Reverse Attention Network (CaraNet) to improve the segmentation performance on small objects compared with several recent state-of-the-art models. CaraNet applies axial reserve attention (ARA) and channel-wise feature pyramid (CFP) module to dig feature information of small medical object. And we evaluate our model by six different measurement metrics. We test our CaraNet on brain tumor (BraTS 2018) and polyp (Kvasir-SEG, CVC-ColonDB, CVC-ClinicDB, CVC-300, and ETIS-LaribPolypDB) segmentation datasets. Our CaraNet achieves the top-rank mean Dice segmentation accuracy, and results show a distinct advantage of CaraNet in the segmentation of small medical objects.
comment: arXiv admin note: text overlap with arXiv:2108.07368
☆ Iterative Loop Learning Combining Self-Training and Active Learning for Domain Adaptive Semantic Segmentation
Recently, self-training and active learning have been proposed to alleviate this problem. Self-training can improve model accuracy with massive unlabeled data, but some pseudo labels containing noise would be generated with limited or imbalanced training data. And there will be suboptimal models if human guidance is absent. Active learning can select more effective data to intervene, while the model accuracy can not be improved because the massive unlabeled data are not used. And the probability of querying sub-optimal samples will increase when the domain difference is too large, increasing annotation cost. This paper proposes an iterative loop learning method combining Self-Training and Active Learning (STAL) for domain adaptive semantic segmentation. The method first uses self-training to learn massive unlabeled data to improve model accuracy and provide more accurate selection models for active learning. Secondly, combined with the sample selection strategy of active learning, manual intervention is used to correct the self-training learning. Iterative loop to achieve the best performance with minimal label cost. Extensive experiments show that our method establishes state-of-the-art performance on tasks of GTAV to Cityscapes, SYNTHIA to Cityscapes, improving by 4.9% mIoU and 5.2% mIoU, compared to the previous best method, respectively. Code will be available.
comment: 11 pages,5 figures
☆ Skeleton-based Human Action Recognition via Convolutional Neural Networks (CNN)
Recently, there has been a remarkable increase in the interest towards skeleton-based action recognition within the research community, owing to its various advantageous features, including computational efficiency, representative features, and illumination invariance. Despite this, researchers continue to explore and investigate the most optimal way to represent human actions through skeleton representation and the extracted features. As a result, the growth and availability of human action recognition datasets have risen substantially. In addition, deep learning-based algorithms have gained widespread popularity due to the remarkable advancements in various computer vision tasks. Most state-of-the-art contributions in skeleton-based action recognition incorporate a Graph Neural Network (GCN) architecture for representing the human body and extracting features. Our research demonstrates that Convolutional Neural Networks (CNNs) can attain comparable results to GCN, provided that the proper training techniques, augmentations, and optimizers are applied. Our approach has been rigorously validated, and we have achieved a score of 95% on the NTU-60 dataset
☆ IM-IAD: Industrial Image Anomaly Detection Benchmark in Manufacturing
Image anomaly detection (IAD) is an emerging and vital computer vision task in industrial manufacturing (IM). Recently many advanced algorithms have been published, but their performance deviates greatly. We realize that the lack of actual IM settings most probably hinders the development and usage of these methods in real-world applications. As far as we know, IAD methods are not evaluated systematically. As a result, this makes it difficult for researchers to analyze them because they are designed for different or special cases. To solve this problem, we first propose a uniform IM setting to assess how well these algorithms perform, which includes several aspects, i.e., various levels of supervision (unsupervised vs. semi-supervised), few-shot learning, continual learning, noisy labels, memory usage, and inference speed. Moreover, we skillfully build a comprehensive image anomaly detection benchmark (IM-IAD) that includes 16 algorithms on 7 mainstream datasets with uniform settings. Our extensive experiments (17,017 in total) provide in-depth insights for IAD algorithm redesign or selection under the IM setting. Next, the proposed benchmark IM-IAD gives challenges as well as directions for the future. To foster reproducibility and accessibility, the source code of IM-IAD is uploaded on the website, https://github.com/M-3LAB/IM-IAD.
☆ Hierarchical Disentangled Representation for Invertible Image Denoising and Beyond
Image denoising is a typical ill-posed problem due to complex degradation. Leading methods based on normalizing flows have tried to solve this problem with an invertible transformation instead of a deterministic mapping. However, the implicit bijective mapping is not explored well. Inspired by a latent observation that noise tends to appear in the high-frequency part of the image, we propose a fully invertible denoising method that injects the idea of disentangled learning into a general invertible neural network to split noise from the high-frequency part. More specifically, we decompose the noisy image into clean low-frequency and hybrid high-frequency parts with an invertible transformation and then disentangle case-specific noise and high-frequency components in the latent space. In this way, denoising is made tractable by inversely merging noiseless low and high-frequency parts. Furthermore, we construct a flexible hierarchical disentangling framework, which aims to decompose most of the low-frequency image information while disentangling noise from the high-frequency part in a coarse-to-fine manner. Extensive experiments on real image denoising, JPEG compressed artifact removal, and medical low-dose CT image restoration have demonstrated that the proposed method achieves competing performance on both quantitative metrics and visual quality, with significantly less computational cost.
comment: Technical Report
☆ Inference Time Evidences of Adversarial Attacks for Forensic on Transformers
Vision Transformers (ViTs) are becoming a very popular paradigm for vision tasks as they achieve state-of-the-art performance on image classification. However, although early works implied that this network structure had increased robustness against adversarial attacks, some works argue ViTs are still vulnerable. This paper presents our first attempt toward detecting adversarial attacks during inference time using the network's input and outputs as well as latent features. We design four quantifications (or derivatives) of input, output, and latent vectors of ViT-based models that provide a signature of the inference, which could be beneficial for the attack detection, and empirically study their behavior over clean samples and adversarial samples. The results demonstrate that the quantifications from input (images) and output (posterior probabilities) are promising for distinguishing clean and adversarial samples, while latent vectors offer less discriminative power, though they give some insights on how adversarial perturbations work.
☆ Few-Shot Image-to-Semantics Translation for Policy Transfer in Reinforcement Learning IJCNN2022
We investigate policy transfer using image-to-semantics translation to mitigate learning difficulties in vision-based robotics control agents. This problem assumes two environments: a simulator environment with semantics, that is, low-dimensional and essential information, as the state space, and a real-world environment with images as the state space. By learning mapping from images to semantics, we can transfer a policy, pre-trained in the simulator, to the real world, thereby eliminating real-world on-policy agent interactions to learn, which are costly and risky. In addition, using image-to-semantics mapping is advantageous in terms of the computational efficiency to train the policy and the interpretability of the obtained policy over other types of sim-to-real transfer strategies. To tackle the main difficulty in learning image-to-semantics mapping, namely the human annotation cost for producing a training dataset, we propose two techniques: pair augmentation with the transition function in the simulator environment and active learning. We observed a reduction in the annotation cost without a decline in the performance of the transfer, and the proposed approach outperformed the existing approach without annotation.
comment: The 2022 International Joint Conference on Neural Networks (IJCNN2022)
☆ Continuous Spatiotemporal Transformers
Modeling spatiotemporal dynamical systems is a fundamental challenge in machine learning. Transformer models have been very successful in NLP and computer vision where they provide interpretable representations of data. However, a limitation of transformers in modeling continuous dynamical systems is that they are fundamentally discrete time and space models and thus have no guarantees regarding continuous sampling. To address this challenge, we present the Continuous Spatiotemporal Transformer (CST), a new transformer architecture that is designed for the modeling of continuous systems. This new framework guarantees a continuous and smooth output via optimization in Sobolev space. We benchmark CST against traditional transformers as well as other spatiotemporal dynamics modeling methods and achieve superior performance in a number of tasks on synthetic and real systems, including learning brain dynamics from calcium imaging data.
☆ Design and Implementation of A Soccer Ball Detection System with Multiple Cameras
The detection of small and medium-sized objects in three dimensions has always been a frontier exploration problem. This technology has a very wide application in sports analysis, games, virtual reality, human animation and other fields. The traditional three-dimensional small target detection technology has the disadvantages of high cost, low precision and inconvenience, so it is difficult to apply in practice. With the development of machine learning and deep learning, the technology of computer vision algorithms is becoming more mature. Creating an immersive media experience is considered to be a very important research work in sports. The main work is to explore and solve the problem of football detection under the multiple cameras, aiming at the research and implementation of the live broadcast system of football matches. Using multi cameras detects a target ball and determines its position in three dimension with the occlusion, motion, low illumination of the target object. This paper designed and implemented football detection system under multiple cameras for the detection and capture of targets in real-time matches. The main work mainly consists of three parts, football detector, single camera detection, and multi-cameras detection. The system used bundle adjustment to obtain the three-dimensional position of the target, and the GPU to accelerates data pre-processing and achieve accurate real-time capture of the target. By testing the system, it shows that the system can accurately detect and capture the moving targets in 3D. In addition, the solution in this paper is reusable for large-scale competitions, like basketball and soccer. The system framework can be well transplanted into other similar engineering project systems. It has been put into the market.
comment: 89 pages
☆ Real Estate Property Valuation using Self-Supervised Vision Transformers
The use of Artificial Intelligence (AI) in the real estate market has been growing in recent years. In this paper, we propose a new method for property valuation that utilizes self-supervised vision transformers, a recent breakthrough in computer vision and deep learning. Our proposed algorithm uses a combination of machine learning, computer vision and hedonic pricing models trained on real estate data to estimate the value of a given property. We collected and pre-processed a data set of real estate properties in the city of Boulder, Colorado and used it to train, validate and test our algorithm. Our data set consisted of qualitative images (including house interiors, exteriors, and street views) as well as quantitative features such as the number of bedrooms, bathrooms, square footage, lot square footage, property age, crime rates, and proximity to amenities. We evaluated the performance of our model using metrics such as Root Mean Squared Error (RMSE). Our findings indicate that these techniques are able to accurately predict the value of properties, with a low RMSE. The proposed algorithm outperforms traditional appraisal methods that do not leverage property images and has the potential to be used in real-world applications.
☆ Debiasing Vision-Language Models via Biased Prompts
Machine learning models have been shown to inherit biases from their training datasets, which can be particularly problematic for vision-language foundation models trained on uncurated datasets scraped from the internet. The biases can be amplified and propagated to downstream applications like zero-shot classifiers and text-to-image generative models. In this study, we propose a general approach for debiasing vision-language foundation models by projecting out biased directions in the text embedding. In particular, we show that debiasing only the text embedding with a calibrated projection matrix suffices to yield robust classifiers and fair generative models. The closed-form solution enables easy integration into large-scale pipelines, and empirical results demonstrate that our approach effectively reduces social bias and spurious correlation in both discriminative and generative vision-language models without the need for additional data or training.
☆ NASiam: Efficient Representation Learning using Neural Architecture Search for Siamese Networks
Siamese networks are one of the most trending methods to achieve self-supervised visual representation learning (SSL). Since hand labeling is costly, SSL can play a crucial part by allowing deep learning to train on large unlabeled datasets. Meanwhile, Neural Architecture Search (NAS) is becoming increasingly important as a technique to discover novel deep learning architectures. However, early NAS methods based on reinforcement learning or evolutionary algorithms suffered from ludicrous computational and memory costs. In contrast, differentiable NAS, a gradient-based approach, has the advantage of being much more efficient and has thus retained most of the attention in the past few years. In this article, we present NASiam, a novel approach that uses for the first time differentiable NAS to improve the multilayer perceptron projector and predictor (encoder/predictor pair) architectures inside siamese-networks-based contrastive learning frameworks (e.g., SimCLR, SimSiam, and MoCo) while preserving the simplicity of previous baselines. We crafted a search space designed explicitly for multilayer perceptrons, inside which we explored several alternatives to the standard ReLU activation function. We show that these new architectures allow ResNet backbone convolutional models to learn strong representations efficiently. NASiam reaches competitive performance in both small-scale (i.e., CIFAR-10/CIFAR-100) and large-scale (i.e., ImageNet) image classification datasets while costing only a few GPU hours. We discuss the composition of the NAS-discovered architectures and emit hypotheses on why they manage to prevent collapsing behavior. Our code is available at https://github.com/aheuillet/NASiam.
comment: 8 pages, 6 figures
☆ A Comprehensive Survey of Continual Learning: Theory, Method and Application
To cope with real-world dynamics, an intelligent agent needs to incrementally acquire, update, accumulate, and exploit knowledge throughout its lifetime. This ability, known as continual learning, provides a foundation for AI systems to develop themselves adaptively. In a general sense, continual learning is explicitly limited by catastrophic forgetting, where learning a new task usually results in a dramatic performance drop of the old tasks. Beyond this, increasingly numerous advances have emerged in recent years that largely extend the understanding and application of continual learning. The growing and widespread interest in this direction demonstrates its realistic significance as well as complexity. In this work, we present a comprehensive survey of continual learning, seeking to bridge the basic settings, theoretical foundations, representative methods, and practical applications. Based on existing theoretical and empirical results, we summarize the general objectives of continual learning as ensuring a proper stability-plasticity trade-off and an adequate intra/inter-task generalizability in the context of resource efficiency. Then we provide a state-of-the-art and elaborated taxonomy, extensively analyzing how representative strategies address continual learning, and how they are adapted to particular challenges in various applications. Through an in-depth discussion of continual learning in terms of the current trends, cross-directional prospects and interdisciplinary connections with neuroscience, we believe that such a holistic perspective can greatly facilitate subsequent exploration in this field and beyond.
♻ ☆ Py-Feat: Python Facial Expression Analysis Toolbox
Studying facial expressions is a notoriously difficult endeavor. Recent advances in the field of affective computing have yielded impressive progress in automatically detecting facial expressions from pictures and videos. However, much of this work has yet to be widely disseminated in social science domains such as psychology. Current state of the art models require considerable domain expertise that is not traditionally incorporated into social science training programs. Furthermore, there is a notable absence of user-friendly and open-source software that provides a comprehensive set of tools and functions that support facial expression research. In this paper, we introduce Py-Feat, an open-source Python toolbox that provides support for detecting, preprocessing, analyzing, and visualizing facial expression data. Py-Feat makes it easy for domain experts to disseminate and benchmark computer vision models and also for end users to quickly process, analyze, and visualize face expression data. We hope this platform will facilitate increased use of facial expression data in human behavior research.
♻ ☆ Video Influencers: Unboxing the Mystique
Influencer marketing has become a very popular tool to reach customers. Despite the rapid growth in influencer videos, there has been little research on the effectiveness of their constituent elements in explaining video engagement. We study YouTube influencers and analyze their unstructured video data across text, audio and images using a novel "interpretable deep learning" framework that accomplishes both goals of prediction and interpretation. Our prediction-based approach analyzes unstructured data and finds that "what is said" in words (text) is more influential than "how it is said" in imagery (images) followed by acoustics (audio). Our interpretation-based approach is implemented after completion of model prediction by analyzing the same source of unstructured data to measure importance attributed to the video elements. We eliminate several spurious and confounded relationships, and identify a smaller subset of theory-based relationships. We uncover novel findings that establish distinct effects for measures of shallow and deep engagement which are based on the dual-system framework of human thinking. Our approach is validated using simulated data, and we discuss the learnings from our findings for influencers and brands.
comment: 45 pages, Online Appendix
♻ ☆ Learning useful representations for shifting tasks and distributions
Does the dominant approach to learn representations (as a side effect of optimizing an expected cost for a single training distribution) remain a good approach when we are dealing with multiple distributions? Our thesis is that such scenarios are better served by representations that are richer than those obtained with a single optimization episode. We support this thesis with simple theoretical arguments and with experiments utilizing an apparently na\"{\i}ve ensembling technique: concatenating the representations obtained from multiple training episodes using the same data, model, algorithm, and hyper-parameters, but different random seeds. These independently trained networks perform similarly. Yet, in a number of scenarios involving new distributions, the concatenated representation performs substantially better than an equivalently sized network trained with a single training run. This proves that the representations constructed by multiple training episodes are in fact different. Although their concatenation carries little additional information about the training task under the training distribution, it becomes substantially more informative when tasks or distributions change. Meanwhile, a single training episode is unlikely to yield such a redundant representation because the optimization process has no reason to accumulate features that do not incrementally improve the training performance.
comment: 20 pages
♻ ☆ Rethinking Lightweight Convolutional Neural Networks for Efficient and High-quality Pavement Crack Detection
Pixel-level road crack detection has always been a challenging task in intelligent transportation systems. Due to the external environments, such as weather, light, and other factors, pavement cracks often present low contrast, poor continuity, and different sizes in length and width. However, most of the existing studies pay less attention to crack data under different situations. Meanwhile, recent algorithms based on deep convolutional neural networks (DCNNs) have promoted the development of cutting-edge models for crack detection. Nevertheless, they usually focus on complex models for good performance, but ignore detection efficiency in practical applications. In this article, to address the first issue, we collected two new databases (i.e. Rain365 and Sun520) captured in rainy and sunny days respectively, which enrich the data of the open source community. For the second issue, we reconsider how to improve detection efficiency with excellent performance, and then propose our lightweight encoder-decoder architecture termed CarNet. Specifically, we introduce a novel olive-shaped structure for the encoder network, a light-weight multi-scale block and a new up-sampling method in the decoder network. Numerous experiments show that our model can better balance detection performance and efficiency compared with previous models. Especially, on the Sun520 dataset, our CarNet significantly advances the state-of-the-art performance with ODS F-score from 0.488 to 0.514. Meanwhile, it does so with an improved detection speed (104 frame per second) which is orders of magnitude faster than some recent DCNNs-based algorithms specially designed for crack detection.
comment: 19 pages, 14 figures, 12 tables
♻ ☆ What I Cannot Predict, I Do Not Understand: A Human-Centered Evaluation Framework for Explainability Methods
A multitude of explainability methods and associated fidelity performance metrics have been proposed to help better understand how modern AI systems make decisions. However, much of the current work has remained theoretical -- without much consideration for the human end-user. In particular, it is not yet known (1) how useful current explainability methods are in practice for more real-world scenarios and (2) how well associated performance metrics accurately predict how much knowledge individual explanations contribute to a human end-user trying to understand the inner-workings of the system. To fill this gap, we conducted psychophysics experiments at scale to evaluate the ability of human participants to leverage representative attribution methods for understanding the behavior of different image classifiers representing three real-world scenarios: identifying bias in an AI system, characterizing the visual strategy it uses for tasks that are too difficult for an untrained non-expert human observer as well as understanding its failure cases. Our results demonstrate that the degree to which individual attribution methods help human participants better understand an AI system varied widely across these scenarios. This suggests a critical need for the field to move past quantitative improvements of current attribution methods towards the development of complementary approaches that provide qualitatively different sources of information to human end-users.
♻ ☆ Normalized Weighting Schemes for Image Interpolation Algorithms
Image interpolation algorithms pervade many modern image processing and analysis applications. However, when their weighting schemes inefficiently generate very unrealistic estimates, they may negatively affect the performance of the end user applications. Therefore, in this work, the author introduced four weighting schemes based on some geometric shapes for digital image interpolation operations. And, the quantity used to express the extent of each shape weight was the normalized area, especially when the sums of areas exceeded a unit square size. The introduced four weighting schemes are based on the minimum side based diameter (MD) of a regular tetragon, hypotenuse based radius (HR), the virtual pixel length based height for the area of the triangle (AT), and the virtual pixel length for hypotenuse based radius for the area of the circle (AC). At the smaller scaling ratio, the image interpolation algorithm based on the HR scheme scored the highest at 66.6 % among non traditional image interpolation algorithms presented. But, at the higher scaling ratio, the AC scheme based image interpolation algorithm scored the highest at 66.6 % among non traditional algorithms presented and, here, its image interpolation quality was generally superior or comparable to the quality of images interpolated by both non traditional and traditional algorithms.
comment: 17 pages, 15 figures, 2 Tables
♻ ☆ Canonical Cortical Graph Neural Networks and its Application for Speech Enhancement in Audio-Visual Hearing Aids
Despite the recent success of machine learning algorithms, most models face drawbacks when considering more complex tasks requiring interaction between different sources, such as multimodal input data and logical time sequences. On the other hand, the biological brain is highly sharpened in this sense, empowered to automatically manage and integrate such streams of information. In this context, this work draws inspiration from recent discoveries in brain cortical circuits to propose a more biologically plausible self-supervised machine learning approach. This combines multimodal information using intra-layer modulations together with Canonical Correlation Analysis, and a memory mechanism to keep track of temporal data, the overall approach termed Canonical Cortical Graph Neural networks. This is shown to outperform recent state-of-the-art models in terms of clean audio reconstruction and energy efficiency for a benchmark audio-visual speech dataset. The enhanced performance is demonstrated through a reduced and smother neuron firing rate distribution. suggesting that the proposed model is amenable for speech enhancement in future audio-visual hearing aid devices.
♻ ☆ Swarm-SLAM : Sparse Decentralized Collaborative Simultaneous Localization and Mapping Framework for Multi-Robot Systems
Collaborative Simultaneous Localization And Mapping (C-SLAM) is a vital component for successful multi-robot operations in environments without an external positioning system, such as indoors, underground or underwater. In this paper, we introduce Swarm-SLAM, an open-source C-SLAM system that is designed to be scalable, flexible, decentralized, and sparse, which are all key properties in swarm robotics. Our system supports inertial, lidar, stereo, and RGB-D sensing, and it includes a novel inter-robot loop closure prioritization technique that reduces communication and accelerates convergence. We evaluated our ROS-2 implementation on five different datasets, and in a real-world experiment with three robots communicating through an ad-hoc network. Our code is publicly available: https://github.com/MISTLab/Swarm-SLAM
comment: Code: https://github.com/MISTLab/Swarm-SLAM
♻ ☆ LSDM: Long-Short Diffeomorphic Motion for Weakly-Supervised Ultrasound Landmark Tracking
Accurate tracking of an anatomical landmark over time has been of high interests for disease assessment such as minimally invasive surgery and tumor radiation therapy. Ultrasound imaging is a promising modality benefiting from low-cost and real-time acquisition. However, generating a precise landmark tracklet is very challenging, as attempts can be easily distorted by different interference such as landmark deformation, visual ambiguity and partial observation. In this paper, we propose a long-short diffeomorphic motion network, which is a multi-task framework with a learnable deformation prior to search for the plausible deformation of landmark. Specifically, we design a novel diffeomorphism representation in both long and short temporal domains for delineating motion margins and reducing long-term cumulative tracking errors. To further mitigate local anatomical ambiguity, we propose an expectation maximisation motion alignment module to iteratively optimize both long and short deformation, aligning to the same directional and spatial representation. The proposed multi-task system can be trained in a weakly-supervised manner, which only requires few landmark annotations for tracking and zero annotation for long-short deformation learning. We conduct extensive experiments on two ultrasound landmark tracking datasets. Experimental results show that our proposed method can achieve better or competitive landmark tracking performance compared with other state-of-the-art tracking methods, with a strong generalization capability across different scanner types and different ultrasound modalities.
♻ ☆ PromptMix: Text-to-image diffusion models enhance the performance of lightweight networks
Many deep learning tasks require annotations that are too time consuming for human operators, resulting in small dataset sizes. This is especially true for dense regression problems such as crowd counting which requires the location of every person in the image to be annotated. Techniques such as data augmentation and synthetic data generation based on simulations can help in such cases. In this paper, we introduce PromptMix, a method for artificially boosting the size of existing datasets, that can be used to improve the performance of lightweight networks. First, synthetic images are generated in an end-to-end data-driven manner, where text prompts are extracted from existing datasets via an image captioning deep network, and subsequently introduced to text-to-image diffusion models. The generated images are then annotated using one or more high-performing deep networks, and mixed with the real dataset for training the lightweight network. By extensive experiments on five datasets and two tasks, we show that PromptMix can significantly increase the performance of lightweight networks by up to 26%.
♻ ☆ Apples and Oranges? Assessing Image Quality over Content Recognition
Image recognition and quality assessment are two important viewing tasks, while potentially following different visual mechanisms. This paper investigates if the two tasks can be performed in a multitask learning manner. A sequential spatial-channel attention module is proposed to simulate the visual attention and contrast sensitivity mechanisms that are crucial for content recognition and quality assessment. Spatial attention is shared between content recognition and quality assessment, while channel attention is solely for quality assessment. Such attention module is integrated into Transformer to build a uniform model for the two viewing tasks. The experimental results have demonstrated that the proposed uniform model can achieve promising performance for both quality assessment and content recognition tasks.
♻ ☆ Disentangling Model Multiplicity in Deep Learning
Model multiplicity is a well-known but poorly understood phenomenon that undermines the generalisation guarantees of machine learning models. It appears when two models with similar training-time performance differ in their predictions and real-world performance characteristics. This observed 'predictive' multiplicity (PM) also implies elusive differences in the internals of the models, their 'representational' multiplicity (RM). We introduce a conceptual and experimental setup for analysing RM by measuring activation similarity via singular vector canonical correlation analysis (SVCCA). We show that certain differences in training methods systematically result in larger RM than others and evaluate RM and PM over a finite sample as predictors for generalizability. We further correlate RM with PM measured by the variance in i.i.d. and out-of-distribution test predictions in four standard image data sets. Finally, instead of attempting to eliminate RM, we call for its systematic measurement and maximal exposure.
comment: 13 pages, 6 figures
♻ ☆ FreeMatch: Self-adaptive Thresholding for Semi-supervised Learning ICLR 2023
Semi-supervised Learning (SSL) has witnessed great success owing to the impressive performances brought by various methods based on pseudo labeling and consistency regularization. However, we argue that existing methods might fail to utilize the unlabeled data more effectively since they either use a pre-defined / fixed threshold or an ad-hoc threshold adjusting scheme, resulting in inferior performance and slow convergence. We first analyze a motivating example to obtain intuitions on the relationship between the desirable threshold and model's learning status. Based on the analysis, we hence propose FreeMatch to adjust the confidence threshold in a self-adaptive manner according to the model's learning status. We further introduce a self-adaptive class fairness regularization penalty to encourage the model for diverse predictions during the early training stage. Extensive experiments indicate the superiority of FreeMatch especially when the labeled data are extremely rare. FreeMatch achieves 5.78%, 13.59%, and 1.28% error rate reduction over the latest state-of-the-art method FlexMatch on CIFAR-10 with 1 label per class, STL-10 with 4 labels per class, and ImageNet with 100 labels per class, respectively. Moreover, FreeMatch can also boost the performance of imbalanced SSL. The codes can be found at https://github.com/microsoft/Semi-supervised-learning.
comment: Accepted by ICLR 2023. Code: https://github.com/microsoft/Semi-supervised-learning
♻ ☆ Spectral Maps for Learning on Subgraphs
In graph learning, maps between graphs and their subgraphs frequently arise. For instance, when coarsening or rewiring operations are present along the pipeline, one needs to keep track of the corresponding nodes between the original and modified graphs. Classically, these maps are represented as binary node-to-node correspondence matrices and used as-is to transfer node-wise features between the graphs. In this paper, we argue that simply changing this map representation can bring notable benefits to graph learning tasks. Drawing inspiration from recent progress in geometry processing, we introduce a spectral representation for maps that is easy to integrate into existing graph learning models. This spectral representation is a compact and straightforward plug-in replacement and is robust to topological changes of the graphs. Remarkably, the representation exhibits structural properties that make it interpretable, drawing an analogy with recent results on smooth manifolds. We demonstrate the benefits of incorporating spectral maps in graph learning pipelines, addressing scenarios where a node-to-node map is not well defined, or in the absence of exact isomorphism. Our approach bears practical benefits in knowledge distillation and hierarchical learning, where we show comparable or improved performance at a fraction of the computational cost.
♻ ☆ Mind the Gap: Polishing Pseudo labels for Accurate Semi-supervised Object Detection AAAI
Exploiting pseudo labels (e.g., categories and bounding boxes) of unannotated objects produced by a teacher detector have underpinned much of recent progress in semi-supervised object detection (SSOD). However, due to the limited generalization capacity of the teacher detector caused by the scarce annotations, the produced pseudo labels often deviate from ground truth, especially those with relatively low classification confidences, thus limiting the generalization performance of SSOD. To mitigate this problem, we propose a dual pseudo-label polishing framework for SSOD. Instead of directly exploiting the pseudo labels produced by the teacher detector, we take the first attempt at reducing their deviation from ground truth using dual polishing learning, where two differently structured polishing networks are elaborately developed and trained using synthesized paired pseudo labels and the corresponding ground truth for categories and bounding boxes on the given annotated objects, respectively. By doing this, both polishing networks can infer more accurate pseudo labels for unannotated objects through sufficiently exploiting their context knowledge based on the initially produced pseudo labels, and thus improve the generalization performance of SSOD. Moreover, such a scheme can be seamlessly plugged into the existing SSOD framework for joint end-to-end learning. In addition, we propose to disentangle the polished pseudo categories and bounding boxes of unannotated objects for separate category classification and bounding box regression in SSOD, which enables introducing more unannotated objects during model training and thus further improve the performance. Experiments on both PASCAL VOC and MS COCO benchmarks demonstrate the superiority of the proposed method over existing state-of-the-art baselines.
comment: Accepted by Thirty-Seventh AAAI Conference on Artificial Intelligence (AAAI 2023)
♻ ☆ InstructRL: Simple yet Effective Instruction-Following Agents with Multimodal Transformer
Humans are excellent at understanding language and vision to accomplish a wide range of tasks. In contrast, creating general instruction-following embodied agents remains a difficult challenge. Prior work that uses pure language-only models lack visual grounding, making it difficult to connect language instructions with visual observations. On the other hand, methods that use pre-trained multimodal models typically come with divided language and visual representations, requiring designing specialized network architecture to fuse them together. We propose a simple yet effective model for robots to solve instruction-following tasks in vision-based environments. Our \ours method consists of a multimodal transformer that encodes visual observations and language instructions, and a transformer-based policy that predicts actions based on encoded representations. The multimodal transformer is pre-trained on millions of image-text pairs and natural language text, thereby producing generic cross-modal representations of observations and instructions. The transformer-based policy keeps track of the full history of observations and actions, and predicts actions autoregressively. Despite its simplicity, we show that this unified transformer model outperforms all state-of-the-art pre-trained or trained-from-scratch methods in both single-task and multi-task settings. Our model also shows better model scalability and generalization ability than prior work.
comment: improved results and presentation
♻ ☆ Orbeez-SLAM: A Real-time Monocular Visual SLAM with ORB Features and NeRF-realized Mapping
A spatial AI that can perform complex tasks through visual signals and cooperate with humans is highly anticipated. To achieve this, we need a visual SLAM that easily adapts to new scenes without pre-training and generates dense maps for downstream tasks in real-time. None of the previous learning-based and non-learning-based visual SLAMs satisfy all needs due to the intrinsic limitations of their components. In this work, we develop a visual SLAM named Orbeez-SLAM, which successfully collaborates with implicit neural representation and visual odometry to achieve our goals. Moreover, Orbeez-SLAM can work with the monocular camera since it only needs RGB inputs, making it widely applicable to the real world. Results show that our SLAM is up to 800x faster than the strong baseline with superior rendering outcomes. Code link: https://github.com/MarvinChung/Orbeez-SLAM.
♻ ☆ Recipro-CAM: Gradient-free reciprocal class activation map
Convolutional neural network (CNN) becomes one of the most popular and prominent deep learning architectures for computer vision, but its black box feature hides the internal prediction process. For this reason, AI practitioners have shed light on explainable AI to provide the interpretability of the model behavior. In particular, class activation map (CAM) and Grad-CAM based methods have shown promise results, but they have architectural limitation or gradient computing burden. To resolve these, Score-CAM has been suggested as a gradient-free method, however, it requires more execution time compared to CAM or Grad-CAM based methods. Therefore, we propose a lightweight architecture and gradient free Reciprocal CAM (Recipro-CAM) by spatially masking the extracted feature maps to exploit the correlation between activation maps and network outputs. With the proposed method, we achieved the gains of 1.78 - 3.72% in the ResNet family compared to Score-CAM in Average Drop- Coherence-Complexity (ADCC) metric, excluding the VGG-16 (1.39% drop). In addition, Recipro-CAM exhibits a saliency map generation rate similar to Grad-CAM and approximately 148 times faster than Score-CAM. The source code of Recipro-CAM is available at our data analysis framework.
♻ ☆ SemSup: Semantic Supervision for Simple and Scalable Zero-shot Generalization
Zero-shot learning is the problem of predicting instances over classes not seen during training. One approach to zero-shot learning is providing auxiliary class information to the model. Prior work along this vein have largely used expensive per-instance annotation or singular class-level descriptions, but per-instance descriptions are hard to scale and single class descriptions may not be rich enough. Furthermore, these works have used natural-language descriptions exclusively, simple bi-encoders models, and modality or task-specific methods. These approaches have several limitations: text supervision may not always be available or optimal and bi-encoders may only learn coarse relations between inputs and class descriptions. In this work, we present SemSup, a novel approach that uses (1) a scalable multiple description sampling method which improves performance over single descriptions, (2) alternative description formats such as JSON that are easy to generate and outperform text on certain settings, and (3) hybrid lexical-semantic similarity to leverage fine-grained information in class descriptions. We demonstrate the effectiveness of SemSup across four datasets, two modalities, and three generalization settings. For example, across text and image datasets, SemSup increases unseen class generalization accuracy by 15 points on average compared to the closest baseline.
♻ ☆ Local Contrast and Global Contextual Information Make Infrared Small Object Salient Again
Infrared small object detection (ISOS) aims to segment small objects only covered with several pixels from clutter background in infrared images. It's of great challenge due to: 1) small objects lack of sufficient intensity, shape and texture information; 2) small objects are easily lost in the process where detection models, say deep neural networks, obtain high-level semantic features and image-level receptive fields through successive downsampling. This paper proposes a reliable detection model for ISOS, dubbed UCFNet, which can handle well the two issues. It builds upon central difference convolution (CDC) and fast Fourier convolution (FFC). On one hand, CDC can effectively guide the network to learn the contrast information between small objects and the background, as the contrast information is very essential in human visual system dealing with the ISOS task. On the other hand, FFC can gain image-level receptive fields and extract global information while preventing small objects from being overwhelmed.Experiments on several public datasets demonstrate that our method significantly outperforms the state-of-the-art ISOS models, and can provide useful guidelines for designing better ISOS deep models. Codes will be available soon.
♻ ☆ PaCaNet: A Study on CycleGAN with Transfer Learning for Diversifying Fused Chinese Painting and Calligraphy
AI-Generated Content (AIGC) has recently gained a surge in popularity, powered by its high efficiency and consistency in production, and its capability of being customized and diversified. The cross-modality nature of the representation learning mechanism in most AIGC technology allows for more freedom and flexibility in exploring new types of art that would be impossible in the past. Inspired by the pictogram subset of Chinese characters, we proposed PaCaNet, a CycleGAN-based pipeline for producing novel artworks that fuse two different art types, traditional Chinese painting and calligraphy. In an effort to produce stable and diversified output, we adopted three main technical innovations: 1. Using one-shot learning to increase the creativity of pre-trained models and diversify the content of the fused images. 2. Controlling the preference over generated Chinese calligraphy by freezing randomly sampled parameters in pre-trained models. 3. Using a regularization method to encourage the models to produce images similar to Chinese paintings. Furthermore, we conducted a systematic study to explore the performance of PaCaNet in diversifying fused Chinese painting and calligraphy, which showed satisfying results. In conclusion, we provide a new direction of creating arts by fusing the visual information in paintings and the stroke features in Chinese calligraphy. Our approach creates a unique aesthetic experience rooted in the origination of Chinese hieroglyph characters. It is also a unique opportunity to delve deeper into traditional artwork and, in doing so, to create a meaningful impact on preserving and revitalizing traditional heritage.
♻ ☆ Adaptively Weighted Data Augmentation Consistency Regularization for Robust Optimization under Concept Shift
Concept shift is a prevailing problem in natural tasks like medical image segmentation where samples usually come from different subpopulations with variant correlations between features and labels. One common type of concept shift in medical image segmentation is the "information imbalance" between label-sparse samples with few (if any) segmentation labels and label-dense samples with plentiful labeled pixels. Existing distributionally robust algorithms have focused on adaptively truncating/down-weighting the "less informative" (i.e., label-sparse in our context) samples. To exploit data features of label-sparse samples more efficiently, we propose an adaptively weighted online optimization algorithm -- AdaWAC -- to incorporate data augmentation consistency regularization in sample reweighting. Our method introduces a set of trainable weights to balance the supervised loss and unsupervised consistency regularization of each sample separately. At the saddle point of the underlying objective, the weights assign label-dense samples to the supervised loss and label-sparse samples to the unsupervised consistency regularization. We provide a convergence guarantee by recasting the optimization as online mirror descent on a saddle point problem. Our empirical results demonstrate that AdaWAC not only enhances the segmentation performance and sample efficiency but also improves the robustness to concept shift on various medical image segmentation tasks with different UNet-style backbones.
♻ ☆ Dynamic Storyboard Generation in an Engine-based Virtual Environment for Video Production
Amateurs working on mini-films and short-form videos usually spend lots of time and effort on the multi-round complicated process of setting and adjusting scenes, plots, and cameras to deliver satisfying video shots. We present Virtual Dynamic Storyboard (VDS) to allow users storyboarding shots in virtual environments, where the filming staff can easily test the settings of shots before the actual filming. VDS runs on a "propose-simulate-discriminate" mode: Given a formatted story script and a camera script as input, it generates several character animation and camera movement proposals following predefined story and cinematic rules to allow an off-the-shelf simulation engine to render videos. To pick up the top-quality dynamic storyboard from the candidates, we equip it with a shot ranking discriminator based on shot quality criteria learned from professional manual-created data. VDS is comprehensively validated via extensive experiments and user studies, demonstrating its efficiency, effectiveness, and great potential in assisting amateur video production.
comment: Project page: https://virtualfilmstudio.github.io/
♻ ☆ Deep Convolutional Pooling Transformer for Deepfake Detection
Recently, Deepfake has drawn considerable public attention due to security and privacy concerns in social media digital forensics. As the wildly spreading Deepfake videos on the Internet become more realistic, traditional detection techniques have failed in distinguishing between real and fake. Most existing deep learning methods mainly focus on local features and relations within the face image using convolutional neural networks as a backbone. However, local features and relations are insufficient for model training to learn enough general information for Deepfake detection. Therefore, the existing Deepfake detection methods have reached a bottleneck to further improve the detection performance. To address this issue, we propose a deep convolutional Transformer to incorporate the decisive image features both locally and globally. Specifically, we apply convolutional pooling and re-attention to enrich the extracted features and enhance efficacy. Moreover, we employ the barely discussed image keyframes in model training for performance improvement and visualize the feature quantity gap between the key and normal image frames caused by video compression. We finally illustrate the transferability with extensive experiments on several Deepfake benchmark datasets. The proposed solution consistently outperforms several state-of-the-art baselines on both within- and cross-dataset experiments.
comment: 13 pages for peer review
♻ ☆ Can we achieve robustness from data alone?
We introduce a meta-learning algorithm for adversarially robust classification. The proposed method tries to be as model agnostic as possible and optimizes a dataset prior to its deployment in a machine learning system, aiming to effectively erase its non-robust features. Once the dataset has been created, in principle no specialized algorithm (besides standard gradient descent) is needed to train a robust model. We formulate the data optimization procedure as a bi-level optimization problem on kernel regression, with a class of kernels that describe infinitely wide neural nets (Neural Tangent Kernels). We present extensive experiments on standard computer vision benchmarks using a variety of different models, demonstrating the effectiveness of our method, while also pointing out its current shortcomings. In parallel, we revisit prior work that also focused on the problem of data optimization for robust classification \citep{Ily+19}, and show that being robust to adversarial attacks after standard (gradient descent) training on a suitable dataset is more challenging than previously thought.
♻ ☆ ERA-Solver: Error-Robust Adams Solver for Fast Sampling of Diffusion Probabilistic Models
Though denoising diffusion probabilistic models (DDPMs) have achieved remarkable generation results, the low sampling efficiency of DDPMs still limits further applications. Since DDPMs can be formulated as diffusion ordinary differential equations (ODEs), various fast sampling methods can be derived from solving diffusion ODEs. However, we notice that previous sampling methods with fixed analytical form are not robust with the error in the noise estimated from pretrained diffusion models. In this work, we construct an error-robust Adams solver (ERA-Solver), which utilizes the implicit Adams numerical method that consists of a predictor and a corrector. Different from the traditional predictor based on explicit Adams methods, we leverage a Lagrange interpolation function as the predictor, which is further enhanced with an error-robust strategy to adaptively select the Lagrange bases with lower error in the estimated noise. Experiments on Cifar10, LSUN-Church, and LSUN-Bedroom datasets demonstrate that our proposed ERA-Solver achieves 5.14, 9.42, and 9.69 Fenchel Inception Distance (FID) for image generation, with only 10 network evaluations.
comment: Typo Error
♻ ☆ Diffusion Denoising for Low-Dose-CT Model
Low-dose Computed Tomography (LDCT) reconstruction is an important task in medical image analysis. Recent years have seen many deep learning based methods, proved to be effective in this area. However, these methods mostly follow a supervised architecture, which needs paired CT image of full dose and quarter dose, and the solution is highly dependent on specific measurements. In this work, we introduce Denoising Diffusion LDCT Model, dubbed as DDLM, generating noise-free CT image using conditioned sampling. DDLM uses pretrained model, and need no training nor tuning process, thus our proposal is in unsupervised manner. Experiments on LDCT images have shown comparable performance of DDLM using less inference time, surpassing other state-of-the-art methods, proving both accurate and efficient. Implementation code will be set to public soon.
♻ ☆ Small Moving Object Detection Algorithm in Surveillance Video Based on Motion Information
A Small Moving Object Detection algorithm Based on Motion Information (SMOD-BMI) is proposed to detect small moving objects with a low Signal-to-Noise Ratio (SNR) in surveillance video. Firstly, a ConvLSTM-PAN model structure is designed to capture suspicious small moving objects, in which the Convolutional Long and Short Time Memory (ConvLSTM) network aggregated the Spatio-temporal features of the adjacent multi-frame small moving object and the Path Aggregation Network (PAN) located the suspicious small moving objects. Then, an object tracking algorithm is used to track suspicious small objects and calculate their Motion Range (MR). At the same time, the size of the MR of the suspicious small moving object is adjusted adaptively according to its speed of movement (specifically, if the object moves slowly, its MR will be expanded according to the speed of the object to ensure the necessary environmental information of the object). Adaptive Spatio-temporal Cubes (ASt-Cubes) of the small moving objects are generated to ensure that the SNR of the moving objects is improved, and the necessary environmental information is retained adaptively. Finally, a LightWeight U-Shape Net (LW-USN) based on ASt-Cubes is designed to detect small moving objects, which rejects false detections and returns the position of small moving objects. This paper uses the bird in the surveillance video as the experimental data set to verify the algorithm's performance. The experimental results show that the proposed small moving object detection method based on motion information in surveillance video can effectively reduce the missed and false detection rate of small moving objects.
♻ ☆ Fast Sampling of Diffusion Models via Operator Learning
Diffusion models have found widespread adoption in various areas. However, their sampling process is slow because it requires hundreds to thousands of network evaluations to emulate a continuous process defined by differential equations. In this work, we use neural operators, an efficient method to solve the probability flow differential equations, to accelerate the sampling process of diffusion models. Compared to other fast sampling methods that have a sequential nature, we are the first to propose parallel decoding method that generates images with only one model forward pass. We propose \textit{diffusion model sampling with neural operator} (DSNO) that maps the initial condition, i.e., Gaussian distribution, to the continuous-time solution trajectory of the reverse diffusion process. To model the temporal correlations along the trajectory, we introduce temporal convolution layers that are parameterized in the Fourier space into the given diffusion model backbone. We show our method achieves state-of-the-art FID of 4.12 for CIFAR-10 and 8.35 for ImageNet-64 in the one-model-evaluation setting.
♻ ☆ Improving Across-Dataset Brain Tissue Segmentation Using Transformer
Brain tissue segmentation has demonstrated great utility in quantifying MRI data through Voxel-Based Morphometry and highlighting subtle structural changes associated with various conditions within the brain. However, manual segmentation is highly labor-intensive, and automated approaches have struggled due to properties inherent to MRI acquisition, leaving a great need for an effective segmentation tool. Despite the recent success of deep convolutional neural networks (CNNs) for brain tissue segmentation, many such solutions do not generalize well to new datasets, which is critical for a reliable solution. Transformers have demonstrated success in natural image segmentation and have recently been applied to 3D medical image segmentation tasks due to their ability to capture long-distance relationships in the input where the local receptive fields of CNNs struggle. This study introduces a novel CNN-Transformer hybrid architecture designed for brain tissue segmentation. We validate our model's performance across four multi-site T1w MRI datasets, covering different vendors, field strengths, scan parameters, time points, and neuropsychiatric conditions. In all situations, our model achieved the greatest generality and reliability. Out method is inherently robust and can serve as a valuable tool for brain-related T1w MRI studies. The code for the TABS network is available at: https://github.com/raovish6/TABS.
♻ ☆ ImpressLearn: Continual Learning via Combined Task Impressions
This work proposes a new method to sequentially train deep neural networks on multiple tasks without suffering catastrophic forgetting, while endowing it with the capability to quickly adapt to unseen tasks. Starting from existing work on network masking (Wortsman et al., 2020), we show that simply learning a linear combination of a small number of task-specific supermasks (impressions) on a randomly initialized backbone network is sufficient to both retain accuracy on previously learned tasks, as well as achieve high accuracy on unseen tasks. In contrast to previous methods, we do not require to generate dedicated masks or contexts for each new task, instead leveraging transfer learning to keep per-task parameter overhead small. Our work illustrates the power of linearly combining individual impressions, each of which fares poorly in isolation, to achieve performance comparable to a dedicated mask. Moreover, even repeated impressions from the same task (homogeneous masks), when combined, can approach the performance of heterogeneous combinations if sufficiently many impressions are used. Our approach scales more efficiently than existing methods, often requiring orders of magnitude fewer parameters and can function without modification even when task identity is missing. In addition, in the setting where task labels are not given at inference, our algorithm gives an often favorable alternative to the one-shot procedure used by Wortsman et al., 2020. We evaluate our method on a number of well-known image classification datasets and network architectures.
♻ ☆ HCR-Net: A deep learning based script independent handwritten character recognition network
Despite being studied extensively for a few decades, handwritten character recognition (HCR) is still considered a challenging learning problem in pattern recognition, and there is very limited research on script independent models. This is mainly because of similarity in structure of characters, different handwriting styles, noisy datasets, diversity of scripts, focus of the conventional research on handcrafted feature extraction techniques, and unavailability of public datasets and code-repositories to reproduce the results. On the other hand, deep learning has witnessed huge success in different areas of pattern recognition, including HCR, and provides an end-to-end learning. However, deep learning techniques are computationally expensive, need large amount of data for training and have been developed for specific scripts only. To address the above limitations, we have proposed a novel generic deep learning architecture for script independent handwritten character recognition, called HCR-Net. HCR-Net is based on a novel transfer learning approach for HCR, which partly utilizes feature extraction layers of a pre-trained network. Due to transfer learning and image-augmentation, HCR-Net provides faster and computationally efficient training, better performance and better generalizations, and can work with small datasets. HCR-Net is extensively evaluated on 40 publicly available datasets of Bangla, Punjabi, Hindi, English, Swedish, Urdu, Farsi, Tibetan, Kannada, Malayalam, Telugu, Marathi, Nepali and Arabic languages, and established 26 new benchmark results while performed close to the best results in the rest cases. HCR-Net showed performance improvements up to 11% against the existing results and achieved a fast convergence rate showing up to 99% of final performance in the very first epoch. HCR-Net significantly outperformed the state-of-the-art transfer learning techniques...
comment: 23 pages (double-column), 6 figures, 16 tables (under review) -- revised version
Information Retrieval 10
☆ Quantum contextual bandits and recommender systems for quantum data
We study a recommender system for quantum data using the linear contextual bandit framework. In each round, a learner receives an observable (the context) and has to recommend from a finite set of unknown quantum states (the actions) which one to measure. The learner has the goal of maximizing the reward in each round, that is the outcome of the measurement on the unknown state. Using this model we formulate the low energy quantum state recommendation problem where the context is a Hamiltonian and the goal is to recommend the state with the lowest energy. For this task, we study two families of contexts: the Ising model and a generalized cluster model. We observe that if we interpret the actions as different phases of the models then the recommendation is done by classifying the correct phase of the given Hamiltonian and the strategy can be interpreted as an online quantum phase classifier.
comment: 15 pages, 9 figures
☆ An Analysis of Classification Approaches for Hit Song Prediction using Engineered Metadata Features with Lyrics and Audio Features
Hit song prediction, one of the emerging fields in music information retrieval (MIR), remains a considerable challenge. Being able to understand what makes a given song a hit is clearly beneficial to the whole music industry. Previous approaches to hit song prediction have focused on using audio features of a record. This study aims to improve the prediction result of the top 10 hits among Billboard Hot 100 songs using more alternative metadata, including song audio features provided by Spotify, song lyrics, and novel metadata-based features (title topic, popularity continuity and genre class). Five machine learning approaches are applied, including: k-nearest neighbours, Naive Bayes, Random Forest, Logistic Regression and Multilayer Perceptron. Our results show that Random Forest (RF) and Logistic Regression (LR) with all features (including novel features, song audio features and lyrics features) outperforms other models, achieving 89.1% and 87.2% accuracy, and 0.91 and 0.93 AUC, respectively. Our findings also demonstrate the utility of our novel music metadata features, which contributed most to the models' discriminative performance.
☆ Large Music Recommendation Studies for Small Teams
Running live music recommendation studies without direct industry partnerships can be a prohibitively daunting task, especially for small teams. In order to help future researchers interested in such evaluations, we present a number of struggles we faced in the process of generating our own such evaluation system alongside potential solutions. These problems span the topics of users, data, computation, and application architecture.
☆ A Counterfactual Collaborative Session-based Recommender System
Most session-based recommender systems (SBRSs) focus on extracting information from the observed items in the current session of a user to predict a next item, ignoring the causes outside the session (called outer-session causes, OSCs) that influence the user's selection of items. However, these causes widely exist in the real world, and few studies have investigated their role in SBRSs. In this work, we analyze the causalities and correlations of the OSCs in SBRSs from the perspective of causal inference. We find that the OSCs are essentially the confounders in SBRSs, which leads to spurious correlations in the data used to train SBRS models. To address this problem, we propose a novel SBRS framework named COCO-SBRS (COunterfactual COllaborative Session-Based Recommender Systems) to learn the causality between OSCs and user-item interactions in SBRSs. COCO-SBRS first adopts a self-supervised approach to pre-train a recommendation model by designing pseudo-labels of causes for each user's selection of the item in data to guide the training process. Next, COCO-SBRS adopts counterfactual inference to recommend items based on the outputs of the pre-trained recommendation model considering the causalities to alleviate the data sparsity problem. As a result, COCO-SBRS can learn the causalities in data, preventing the model from learning spurious correlations. The experimental results of our extensive experiments conducted on three real-world datasets demonstrate the superiority of our proposed framework over ten representative SBRSs.
comment: accepted by the ACM WebConf 2023
☆ Machine Translation Impact in E-commerce Multilingual Search EMNLP 2022
Previous work suggests that performance of cross-lingual information retrieval correlates highly with the quality of Machine Translation. However, there may be a threshold beyond which improving query translation quality yields little or no benefit to further improve the retrieval performance. This threshold may depend upon multiple factors including the source and target languages, the existing MT system quality and the search pipeline. In order to identify the benefit of improving an MT system for a given search pipeline, we investigate the sensitivity of retrieval quality to the presence of different levels of MT quality using experimental datasets collected from actual traffic. We systematically improve the performance of our MT systems quality on language pairs as measured by MT evaluation metrics including Bleu and Chrf to determine their impact on search precision metrics and extract signals that help to guide the improvement strategies. Using this information we develop techniques to compare query translations for multiple language pairs and identify the most promising language pairs to invest and improve.
comment: Accepted by EMNLP 2022 (Industry Track)
☆ In-Context Retrieval-Augmented Language Models
Retrieval-Augmented Language Modeling (RALM) methods, that condition a language model (LM) on relevant documents from a grounding corpus during generation, have been shown to significantly improve language modeling while also providing a natural source attribution mechanism. Existing RALM approaches focus on modifying the LM architecture in order to facilitate the incorporation of external information, significantly complicating deployment. This paper proposes an under-explored alternative, which we dub In-Context RALM: leaving the LM architecture unchanged and prepending grounding documents to the input. We show that in-context RALM which uses off-the-shelf general purpose retrievers provides surprisingly large LM gains across model sizes and diverse corpora. We also demonstrate that the document retrieval and ranking mechanism can be specialized to the RALM setting to further boost performance. We conclude that in-context RALM has considerable potential to increase the prevalence of LM grounding, particularly in settings where a pretrained LM must be used without modification or even via API access. To that end, we make our code publicly available.
♻ ☆ FE-TCM: Filter-Enhanced Transformer Click Model for Web Search
Constructing click models and extracting implicit relevance feedback information from the interaction between users and search engines are very important to improve the ranking of search results. Using neural network to model users' click behaviors has become one of the effective methods to construct click models. In this paper, We use Transformer as the backbone network of feature extraction, add filter layer innovatively, and propose a new Filter-Enhanced Transformer Click Model (FE-TCM) for web search. Firstly, in order to reduce the influence of noise on user behavior data, we use the learnable filters to filter log noise. Secondly, following the examination hypothesis, we model the attraction estimator and examination predictor respectively to output the attractiveness scores and examination probabilities. A novel transformer model is used to learn the deeper representation among different features. Finally, we apply the combination functions to integrate attractiveness scores and examination probabilities into the click prediction. From our experiments on two real-world session datasets, it is proved that FE-TCM outperforms the existing click models for the click prediction.
♻ ☆ CSDR-BERT: a pre-trained scientific dataset match model for Chinese Scientific Dataset Retrieval
As the number of open and shared scientific datasets on the Internet increases under the open science movement, efficiently retrieving these datasets is a crucial task in information retrieval (IR) research. In recent years, the development of large models, particularly the pre-training and fine-tuning paradigm, which involves pre-training on large models and fine-tuning on downstream tasks, has provided new solutions for IR match tasks. In this study, we use the original BERT token in the embedding layer, improve the Sentence-BERT model structure in the model layer by introducing the SimCSE and K-Nearest Neighbors method, and use the cosent loss function in the optimization phase to optimize the target output. Our experimental results show that our model outperforms other competing models on both public and self-built datasets through comparative experiments and ablation implementations. This study explores and validates the feasibility and efficiency of pre-training techniques for semantic retrieval of Chinese scientific datasets.
♻ ☆ Enhancing Dyadic Relations with Homogeneous Graphs for Multimodal Recommendation
User interaction data in recommender systems is a form of dyadic relation that reflects the preferences of users with items. Learning the representations of these two discrete sets of objects, users and items, is critical for recommendation. Recent multimodal recommendation models leveraging multimodal features (e.g., images and text descriptions) have been demonstrated to be effective in improving recommendation accuracy. However, state-of-the-art models enhance the dyadic relations between users and items by considering either user-user or item-item relations, leaving the high-order relations of the other side (i.e., users or items) unexplored. Furthermore, we experimentally reveal that the current multimodality fusion methods in the state-of-the-art models may degrade their recommendation performance. That is, without tainting the model architectures, these models can achieve even better recommendation accuracy with uni-modal information. On top of the finding, we propose a model that enhances the dyadic relations by learning Dual RepresentAtions of both users and items via constructing homogeneous Graphs for multimOdal recommeNdation. We name our model as DRAGON. Specifically, DRAGON constructs the user-user graph based on the commonly interacted items and the item-item graph from item multimodal features. It then utilizes graph learning on both the user-item heterogeneous graph and the homogeneous graphs (user-user and item-item) to obtain the dual representations of users and items. To capture information from each modality, DRAGON employs a simple yet effective fusion method, attentive concatenation, to derive the representations of users and items. Extensive experiments on three public datasets and seven baselines show that DRAGON can outperform the strongest baseline by 22.03% on average. Various ablation studies are conducted on DRAGON to validate its effectiveness.
comment: modify the format
♻ ☆ On Cache-Aided Multi-User Private Information Retrieval with Small Caches
In this paper, we propose a scheme for the problem of cache-aided multi-user private information retrieval with small caches, in which $K$ users are connected to $S$ non-colluding databases via shared links. Each database contains a set of $N$ files, and each user has a dedicated cache of size equivalent to the size of $M$ files. All the users want to retrieve a file without revealing their demands to the databases. During off-peak hours, all the users will fill their caches, and when required, users will demand their desired files by cooperatively generating query sets for each database. After receiving the transmissions from databases, all the users should get their desired files using transmitted data and their cache contents. This problem has been studied in [X. Zhang, K. Wan, H. Sun, M. Ji and G. Caire, \tqt{Fundamental limits of cache-aided multiuser private information retrieval}, IEEE Trans. Commun., 2021], in which authors proposed a product design scheme. In this paper, we propose a scheme that gives a better rate for a particular value of $M$ than the product design scheme. We consider a slightly different approach for the placement phase. Instead of a database filling the caches of all users directly, a database will broadcast cache content for all users on a shared link, and then the users will decide unitedly which part of the broadcasted content will be stored in the cache of each user. This variation facilitates maintaining the privacy constraint at a reduced rate.
comment: 35 pages, 7 tables and 3 figures. Two figures have been added
Machine Learning 156
☆ PADL: Language-Directed Physics-Based Character Control
Developing systems that can synthesize natural and life-like motions for simulated characters has long been a focus for computer animation. But in order for these systems to be useful for downstream applications, they need not only produce high-quality motions, but must also provide an accessible and versatile interface through which users can direct a character's behaviors. Natural language provides a simple-to-use and expressive medium for specifying a user's intent. Recent breakthroughs in natural language processing (NLP) have demonstrated effective use of language-based interfaces for applications such as image generation and program synthesis. In this work, we present PADL, which leverages recent innovations in NLP in order to take steps towards developing language-directed controllers for physics-based character animation. PADL allows users to issue natural language commands for specifying both high-level tasks and low-level skills that a character should perform. We present an adversarial imitation learning approach for training policies to map high-level language commands to low-level controls that enable a character to perform the desired task and skill specified by a user's commands. Furthermore, we propose a multi-task aggregation method that leverages a language-based multiple-choice question-answering approach to determine high-level task objectives from language commands. We show that our framework can be applied to effectively direct a simulated humanoid character to perform a diverse array of complex motor skills.
☆ Mathematical Capabilities of ChatGPT
We investigate the mathematical capabilities of ChatGPT by testing it on publicly available datasets, as well as hand-crafted ones, and measuring its performance against other models trained on a mathematical corpus, such as Minerva. We also test whether ChatGPT can be a useful assistant to professional mathematicians by emulating various use cases that come up in the daily professional activities of mathematicians (question answering, theorem searching). In contrast to formal mathematics, where large databases of formal proofs are available (e.g., the Lean Mathematical Library), current datasets of natural-language mathematics, used to benchmark language models, only cover elementary mathematics. We address this issue by introducing a new dataset: GHOSTS. It is the first natural-language dataset made and curated by working researchers in mathematics that (1) aims to cover graduate-level mathematics and (2) provides a holistic overview of the mathematical capabilities of language models. We benchmark ChatGPT on GHOSTS and evaluate performance against fine-grained criteria. We make this new dataset publicly available to assist a community-driven comparison of ChatGPT with (future) large language models in terms of advanced mathematical comprehension. We conclude that contrary to many positive reports in the media (a potential case of selection bias), ChatGPT's mathematical abilities are significantly below those of an average mathematics graduate student. Our results show that ChatGPT often understands the question but fails to provide correct solutions. Hence, if your goal is to use it to pass a university exam, you would be better off copying from your average peer!
comment: The GHOSTS dataset will be available at https://github.com/friederrr/science-GHOSTS
☆ Salient Conditional Diffusion for Defending Against Backdoor Attacks
We propose a novel algorithm, Salient Conditional Diffusion (Sancdifi), a state-of-the-art defense against backdoor attacks. Sancdifi uses a denoising diffusion probabilistic model (DDPM) to degrade an image with noise and then recover said image using the learned reverse diffusion. Critically, we compute saliency map-based masks to condition our diffusion, allowing for stronger diffusion on the most salient pixels by the DDPM. As a result, Sancdifi is highly effective at diffusing out triggers in data poisoned by backdoor attacks. At the same time, it reliably recovers salient features when applied to clean data. This performance is achieved without requiring access to the model parameters of the Trojan network, meaning Sancdifi operates as a black-box defense.
comment: 12 pages, 5 figures
☆ Learning in POMDPs is Sample-Efficient with Hindsight Observability
POMDPs capture a broad class of decision making problems, but hardness results suggest that learning is intractable even in simple settings due to the inherent partial observability. However, in many realistic problems, more information is either revealed or can be computed during some point of the learning process. Motivated by diverse applications ranging from robotics to data center scheduling, we formulate a \setting (\setshort) as a POMDP where the latent states are revealed to the learner in hindsight and only during training. We introduce new algorithms for the tabular and function approximation settings that are provably sample-efficient with hindsight observability, even in POMDPs that would otherwise be statistically intractable. We give a lower bound showing that the tabular algorithm is optimal in its dependence on latent state and observation cardinalities.
☆ Simplex Random Features
We present Simplex Random Features (SimRFs), a new random feature (RF) mechanism for unbiased approximation of the softmax and Gaussian kernels by geometrical correlation of random projection vectors. We prove that SimRFs provide the smallest possible mean square error (MSE) on unbiased estimates of these kernels among the class of weight-independent geometrically-coupled positive random feature (PRF) mechanisms, substantially outperforming the previously most accurate Orthogonal Random Features at no observable extra cost. We present a more computationally expensive SimRFs+ variant, which we prove is asymptotically optimal in the broader family of weight-dependent geometrical coupling schemes (which permit correlations between random vector directions and norms). In extensive empirical studies, we show consistent gains provided by SimRFs in settings including pointwise kernel estimation, nonparametric classification and scalable Transformers.
☆ Gaussian Noise is Nearly Instance Optimal for Private Unbiased Mean Estimation
We investigate unbiased high-dimensional mean estimators in differential privacy. We consider differentially private mechanisms whose expected output equals the mean of the input dataset, for every dataset drawn from a fixed convex domain $K$ in $\mathbb{R}^d$. In the setting of concentrated differential privacy, we show that, for every input such an unbiased mean estimator introduces approximately at least as much error as a mechanism that adds Gaussian noise with a carefully chosen covariance. This is true when the error is measured with respect to $\ell_p$ error for any $p \ge 2$. We extend this result to local differential privacy, and to approximate differential privacy, but for the latter the error lower bound holds either for a dataset or for a neighboring dataset. We also extend our results to mechanisms that take i.i.d.~samples from a distribution over $K$ and are unbiased with respect to the mean of the distribution.
☆ Benchmarking Large Language Models for News Summarization
Large language models (LLMs) have shown promise for automatic summarization but the reasons behind their successes are poorly understood. By conducting a human evaluation on ten LLMs across different pretraining methods, prompts, and model scales, we make two important observations. First, we find instruction tuning, and not model size, is the key to the LLM's zero-shot summarization capability. Second, existing studies have been limited by low-quality references, leading to underestimates of human performance and lower few-shot and finetuning performance. To better evaluate LLMs, we perform human evaluation over high-quality summaries we collect from freelance writers. Despite major stylistic differences such as the amount of paraphrasing, we find that LMM summaries are judged to be on par with human written summaries.
☆ Interpreting Robustness Proofs of Deep Neural Networks
In recent years numerous methods have been developed to formally verify the robustness of deep neural networks (DNNs). Though the proposed techniques are effective in providing mathematical guarantees about the DNNs behavior, it is not clear whether the proofs generated by these methods are human-interpretable. In this paper, we bridge this gap by developing new concepts, algorithms, and representations to generate human understandable interpretations of the proofs. Leveraging the proposed method, we show that the robustness proofs of standard DNNs rely on spurious input features, while the proofs of DNNs trained to be provably robust filter out even the semantically meaningful features. The proofs for the DNNs combining adversarial and provably robust training are the most effective at selectively filtering out spurious features as well as relying on human-understandable input features.
☆ Grounding Language Models to Images for Multimodal Generation
We propose an efficient method to ground pretrained text-only language models to the visual domain, enabling them to process and generate arbitrarily interleaved image-and-text data. Our method leverages the abilities of language models learnt from large scale text-only pretraining, such as in-context learning and free-form text generation. We keep the language model frozen, and finetune input and output linear layers to enable cross-modality interactions. This allows our model to process arbitrarily interleaved image-and-text inputs, and generate free-form text interleaved with retrieved images. We achieve strong zero-shot performance on grounded tasks such as contextual image retrieval and multimodal dialogue, and showcase compelling interactive abilities. Our approach works with any off-the-shelf language model and paves the way towards an effective, general solution for leveraging pretrained language models in visually grounded settings.
comment: Project page: https://jykoh.com/fromage
☆ Image Shortcut Squeezing: Countering Perturbative Availability Poisons with Compression
Perturbative availability poisoning (PAP) adds small changes to images to prevent their use for model training. Current research adopts the belief that practical and effective approaches to countering such poisons do not exist. In this paper, we argue that it is time to abandon this belief. We present extensive experiments showing that 12 state-of-the-art PAP methods are vulnerable to Image Shortcut Squeezing (ISS), which is based on simple compression. For example, on average, ISS restores the CIFAR-10 model accuracy to $81.73\%$, surpassing the previous best preprocessing-based countermeasures by $37.97\%$ absolute. ISS also (slightly) outperforms adversarial training and has higher generalizability to unseen perturbation norms and also higher efficiency. Our investigation reveals that the property of PAP perturbations depends on the type of surrogate model used for poison generation, and it explains why a specific ISS compression yields the best performance for a specific type of PAP perturbation. We further test stronger, adaptive poisoning, and show it falls short of being an ideal defense against ISS. Overall, our results demonstrate the importance of considering various (simple) countermeasures to ensure the meaningfulness of analysis carried out during the development of availability poisons.
comment: Our code is available at https://github.com/liuzrcc/ImageShortcutSqueezing
☆ A Mathematical Model for Curriculum Learning
Curriculum learning (CL) - training using samples that are generated and presented in a meaningful order - was introduced in the machine learning context around a decade ago. While CL has been extensively used and analysed empirically, there has been very little mathematical justification for its advantages. We introduce a CL model for learning the class of k-parities on d bits of a binary string with a neural network trained by stochastic gradient descent (SGD). We show that a wise choice of training examples, involving two or more product distributions, allows to reduce significantly the computational cost of learning this class of functions, compared to learning under the uniform distribution. We conduct experiments to support our analysis. Furthermore, we show that for another class of functions - namely the `Hamming mixtures' - CL strategies involving a bounded number of product distributions are not beneficial, while we conjecture that CL with unbounded many curriculum steps can learn this class efficiently.
☆ Attend-and-Excite: Attention-Based Semantic Guidance for Text-to-Image Diffusion Models
Recent text-to-image generative models have demonstrated an unparalleled ability to generate diverse and creative imagery guided by a target text prompt. While revolutionary, current state-of-the-art diffusion models may still fail in generating images that fully convey the semantics in the given text prompt. We analyze the publicly available Stable Diffusion model and assess the existence of catastrophic neglect, where the model fails to generate one or more of the subjects from the input prompt. Moreover, we find that in some cases the model also fails to correctly bind attributes (e.g., colors) to their corresponding subjects. To help mitigate these failure cases, we introduce the concept of Generative Semantic Nursing (GSN), where we seek to intervene in the generative process on the fly during inference time to improve the faithfulness of the generated images. Using an attention-based formulation of GSN, dubbed Attend-and-Excite, we guide the model to refine the cross-attention units to attend to all subject tokens in the text prompt and strengthen - or excite - their activations, encouraging the model to generate all subjects described in the text prompt. We compare our approach to alternative approaches and demonstrate that it conveys the desired concepts more faithfully across a range of text prompts.
☆ Complete Neural Networks for Euclidean Graphs
We propose a 2-WL-like geometric graph isomorphism test and prove it is complete when applied to Euclidean Graphs in $\mathbb{R}^3$. We then use recent results on multiset embeddings to devise an efficient geometric GNN model with equivalent separation power. We verify empirically that our GNN model is able to separate particularly challenging synthetic examples, and demonstrate its usefulness for a chemical property prediction problem.
☆ Execution-based Code Generation using Deep Reinforcement Learning
The utilization of programming language (PL) models, pretrained on large-scale code corpora, as a means of automating software engineering processes has demonstrated considerable potential in streamlining various code generation tasks such as code completion, code translation, and program synthesis. However, current approaches mainly rely on supervised fine-tuning objectives borrowed from text generation, neglecting specific sequence-level features of code, including but not limited to compilability as well as syntactic and functional correctness. To address this limitation, we propose PPOCoder, a new framework for code generation that combines pretrained PL models with Proximal Policy Optimization (PPO) deep reinforcement learning and employs execution feedback as the external source of knowledge into the model optimization. PPOCoder is transferable across different code generation tasks and PLs. Extensive experiments on three code generation tasks demonstrate the effectiveness of our proposed approach compared to SOTA methods, improving the success rate of compilation and functional correctness over different PLs. Our code can be found at https://github.com/reddy-lab-code-research/PPOCoder .
☆ Identifying the Hazard Boundary of ML-enabled Autonomous Systems Using Cooperative Co-Evolutionary Search
In Machine Learning (ML)-enabled autonomous systems (MLASs), it is essential to identify the hazard boundary of ML Components (MLCs) in the MLAS under analysis. Given that such boundary captures the conditions in terms of MLC behavior and system context that can lead to hazards, it can then be used to, for example, build a safety monitor that can take any predefined fallback mechanisms at runtime when reaching the hazard boundary. However, determining such hazard boundary for an ML component is challenging. This is due to the space combining system contexts (i.e., scenarios) and MLC behaviors (i.e., inputs and outputs) being far too large for exhaustive exploration and even to handle using conventional metaheuristics, such as genetic algorithms. Additionally, the high computational cost of simulations required to determine any MLAS safety violations makes the problem even more challenging. Furthermore, it is unrealistic to consider a region in the problem space deterministically safe or unsafe due to the uncontrollable parameters in simulations and the non-linear behaviors of ML models (e.g., deep neural networks) in the MLAS under analysis. To address the challenges, we propose MLCSHE (ML Component Safety Hazard Envelope), a novel method based on a Cooperative Co-Evolutionary Algorithm (CCEA), which aims to tackle a high-dimensional problem by decomposing it into two lower-dimensional search subproblems. Moreover, we take a probabilistic view of safe and unsafe regions and define a novel fitness function to measure the distance from the probabilistic hazard boundary and thus drive the search effectively. We evaluate the effectiveness and efficiency of MLCSHE on a complex Autonomous Vehicle (AV) case study. Our evaluation results show that MLCSHE is significantly more effective and efficient compared to a standard genetic algorithm and random search.
☆ Fairness-aware Vision Transformer via Debiased Self-Attention
Vision Transformer (ViT) has recently gained significant interest in solving computer vision (CV) problems due to its capability of extracting informative features and modeling long-range dependencies through the self-attention mechanism. To fully realize the advantages of ViT in real-world applications, recent works have explored the trustworthiness of ViT, including its robustness and explainability. However, another desiderata, fairness has not yet been adequately addressed in the literature. We establish that the existing fairness-aware algorithms (primarily designed for CNNs) do not perform well on ViT. This necessitates the need for developing our novel framework via Debiased Self-Attention (DSA). DSA is a fairness-through-blindness approach that enforces ViT to eliminate spurious features correlated with the sensitive attributes for bias mitigation. Notably, adversarial examples are leveraged to locate and mask the spurious features in the input image patches. In addition, DSA utilizes an attention weights alignment regularizer in the training objective to encourage learning informative features for target prediction. Importantly, our DSA framework leads to improved fairness guarantees over prior works on multiple prediction tasks without compromising target prediction performance
☆ Partitioning Distributed Compute Jobs with Reinforcement Learning and Graph Neural Networks
From natural language processing to genome sequencing, large-scale machine learning models are bringing advances to a broad range of fields. Many of these models are too large to be trained on a single machine, and instead must be distributed across multiple devices. This has motivated the research of new compute and network systems capable of handling such tasks. In particular, recent work has focused on developing management schemes which decide how to allocate distributed resources such that some overall objective, such as minimising the job completion time (JCT), is optimised. However, such studies omit explicit consideration of how much a job should be distributed, usually assuming that maximum distribution is desirable. In this work, we show that maximum parallelisation is sub-optimal in relation to user-critical metrics such as throughput and blocking rate. To address this, we propose PAC-ML (partitioning for asynchronous computing with machine learning). PAC-ML leverages a graph neural network and reinforcement learning to learn how much to partition computation graphs such that the number of jobs which meet arbitrary user-defined JCT requirements is maximised. In experiments with five real deep learning computation graphs on a recently proposed optical architecture across four user-defined JCT requirement distributions, we demonstrate PAC-ML achieving up to 56.2% lower blocking rates in dynamic job arrival settings than the canonical maximum parallelisation strategy used by most prior works.
comment: Pre-print
☆ Improved Algorithms for Multi-period Multi-class Packing Problems with~Bandit~Feedback
We consider the linear contextual multi-class multi-period packing problem~(LMMP) where the goal is to pack items such that the total vector of consumption is below a given budget vector and the total value is as large as possible. We consider the setting where the reward and the consumption vector associated with each action is a class-dependent linear function of the context, and the decision-maker receives bandit feedback. LMMP includes linear contextual bandits with knapsacks and online revenue management as special cases. We establish a new more efficient estimator which guarantees a faster convergence rate, and consequently, a lower regret in such problems. We propose a bandit policy that is a closed-form function of said estimated parameters. When the contexts are non-degenerate, the regret of the proposed policy is sublinear in the context dimension, the number of classes, and the time horizon~$T$ when the budget grows at least as $\sqrt{T}$. We also resolve an open problem posed in Agrawal & Devanur (2016), and extend the result to a multi-class setting. Our numerical experiments clearly demonstrate that the performance of our policy is superior to other benchmarks in the literature.
comment: 42 pages including Appendix
☆ Deep learning-based lung segmentation and automatic regional template in chest X-ray images for pediatric tuberculosis SP
Tuberculosis (TB) is still considered a leading cause of death and a substantial threat to global child health. Both TB infection and disease are curable using antibiotics. However, most children who die of TB are never diagnosed or treated. In clinical practice, experienced physicians assess TB by examining chest X-rays (CXR). Pediatric CXR has specific challenges compared to adult CXR, which makes TB diagnosis in children more difficult. Computer-aided diagnosis systems supported by Artificial Intelligence have shown performance comparable to experienced radiologist TB readings, which could ease mass TB screening and reduce clinical burden. We propose a multi-view deep learning-based solution which, by following a proposed template, aims to automatically regionalize and extract lung and mediastinal regions of interest from pediatric CXR images where key TB findings may be present. Experimental results have shown accurate region extraction, which can be used for further analysis to confirm TB finding presence and severity assessment. Code publicly available at https://github.com/dani-capellan/pTB_LungRegionExtractor.
comment: This work has been accepted at the SPIE Medical Imaging 2023, Image Processing conference
☆ Differentially Private Distributed Bayesian Linear Regression with MCMC
We propose a novel Bayesian inference framework for distributed differentially private linear regression. We consider a distributed setting where multiple parties hold parts of the data and share certain summary statistics of their portions in privacy-preserving noise. We develop a novel generative statistical model for privately shared statistics, which exploits a useful distributional relation between the summary statistics of linear regression. Bayesian estimation of the regression coefficients is conducted mainly using Markov chain Monte Carlo algorithms, while we also provide a fast version to perform Bayesian estimation in one iteration. The proposed methods have computational advantages over their competitors. We provide numerical results on both real and simulated data, which demonstrate that the proposed algorithms provide well-rounded estimation and prediction.
comment: 20 pages, 3 figures, code available at: https://github.com/sinanyildirim/Bayesian_DP_dist_LR
☆ Multicalibration as Boosting for Regression
We study the connection between multicalibration and boosting for squared error regression. First we prove a useful characterization of multicalibration in terms of a ``swap regret'' like condition on squared error. Using this characterization, we give an exceedingly simple algorithm that can be analyzed both as a boosting algorithm for regression and as a multicalibration algorithm for a class H that makes use only of a standard squared error regression oracle for H. We give a weak learning assumption on H that ensures convergence to Bayes optimality without the need to make any realizability assumptions -- giving us an agnostic boosting algorithm for regression. We then show that our weak learning assumption on H is both necessary and sufficient for multicalibration with respect to H to imply Bayes optimality. We also show that if H satisfies our weak learning condition relative to another class C then multicalibration with respect to H implies multicalibration with respect to C. Finally we investigate the empirical performance of our algorithm experimentally using an open source implementation that we make available. Our code repository can be found at https://github.com/Declancharrison/Level-Set-Boosting.
comment: Code available here: https://github.com/Declancharrison/Level-Set-Boosting
☆ Semi-Supervised Classification with Graph Convolutional Kernel Machines
We present a deep Graph Convolutional Kernel Machine (GCKM) for semi-supervised node classification in graphs. First, we introduce an unsupervised kernel machine propagating the node features in a one-hop neighbourhood. Then, we specify a semi-supervised classification kernel machine through the lens of the Fenchel-Young inequality. The deep graph convolutional kernel machine is obtained by stacking multiple shallow kernel machines. After showing that unsupervised and semi-supervised layer corresponds to an eigenvalue problem and a linear system on the aggregated node features, respectively, we derive an efficient end-to-end training algorithm in the dual variables. Numerical experiments demonstrate that our approach is competitive with state-of-the-art graph neural networks for homophilious and heterophilious benchmark datasets. Notably, GCKM achieves superior performance when very few labels are available.
☆ Learning, Fast and Slow: A Goal-Directed Memory-Based Approach for Dynamic Environments
Model-based next state prediction and state value prediction are slow to converge. To address these challenges, we do the following: i) Instead of a neural network, we do model-based planning using a parallel memory retrieval system (which we term the slow mechanism); ii) Instead of learning state values, we guide the agent's actions using goal-directed exploration, by using a neural network to choose the next action given the current state and the goal state (which we term the fast mechanism). The goal-directed exploration is trained online using hippocampal replay of visited states and future imagined states every single time step, leading to fast and efficient training. Empirical studies show that our proposed method has a 92% solve rate across 100 episodes in a dynamically changing grid world, significantly outperforming state-of-the-art actor critic mechanisms such as PPO (54%), TRPO (50%) and A2C (24%). Ablation studies demonstrate that both mechanisms are crucial. We posit that the future of Reinforcement Learning (RL) will be to model goals and sub-goals for various tasks, and plan it out in a goal-directed memory-based approach.
comment: 22 pages
☆ Toward Efficient Gradient-Based Value Estimation
Gradient-based methods for value estimation in reinforcement learning have favorable stability properties, but they are typically much slower than Temporal Difference (TD) learning methods. We study the root causes of this slowness and show that Mean Square Bellman Error (MSBE) is an ill-conditioned loss function in the sense that its Hessian has large condition-number. To resolve the adverse effect of poor conditioning of MSBE on gradient based methods, we propose a low complexity batch-free proximal method that approximately follows the Gauss-Newton direction and is asymptotically robust to parameterization. Our main algorithm, called RANS, is efficient in the sense that it is significantly faster than the residual gradient methods while having almost the same computational complexity, and is competitive with TD on the classic problems that we tested.
☆ Retrosynthetic Planning with Dual Value Networks
Retrosynthesis, which aims to find a route to synthesize a target molecule from commercially available starting materials, is a critical task in drug discovery and materials design. Recently, the combination of ML-based single-step reaction predictors with multi-step planners has led to promising results. However, the single-step predictors are mostly trained offline to optimize the single-step accuracy, without considering complete routes. Here, we leverage reinforcement learning (RL) to improve the single-step predictor, by using a tree-shaped MDP to optimize complete routes while retaining single-step accuracy. Desirable routes should be both synthesizable and of low cost. We propose an online training algorithm, called Planning with Dual Value Networks (PDVN), in which two value networks predict the synthesizability and cost of molecules, respectively. To maintain the single-step accuracy, we design a two-branch network structure for the single-step predictor. On the widely-used USPTO dataset, our PDVN algorithm improves the search success rate of existing multi-step planners (e.g., increasing the success rate from 85.79% to 98.95% for Retro*, and reducing the number of model calls by half while solving 99.47% molecules for RetroGraph). Furthermore, PDVN finds shorter synthesis routes (e.g., reducing the average route length from 5.76 to 4.83 for Retro*, and from 5.63 to 4.78 for RetroGraph).
☆ Multi-fidelity covariance estimation in the log-Euclidean geometry
We introduce a multi-fidelity estimator of covariance matrices that employs the log-Euclidean geometry of the symmetric positive-definite manifold. The estimator fuses samples from a hierarchy of data sources of differing fidelities and costs for variance reduction while guaranteeing definiteness, in contrast with previous approaches. The new estimator makes covariance estimation tractable in applications where simulation or data collection is expensive; to that end, we develop an optimal sample allocation scheme that minimizes the mean-squared error of the estimator given a fixed budget. Guaranteed definiteness is crucial to metric learning, data assimilation, and other downstream tasks. Evaluations of our approach using data from physical applications (heat conduction, fluid dynamics) demonstrate more accurate metric learning and speedups of more than one order of magnitude compared to benchmarks.
comment: 27 pages, 10 figures, code supplement
☆ Archetypal Analysis++: Rethinking the Initialization Strategy
Archetypal analysis is a matrix factorization method with convexity constraints. Due to local minima, a good initialization is essential. Frequently used initialization methods yield either sub-optimal starting points or are prone to get stuck in poor local minima. In this paper, we propose archetypal analysis++ (AA++), a probabilistic initialization strategy for archetypal analysis that sequentially samples points based on their influence on the objective, similar to $k$-means++. In fact, we argue that $k$-means++ already approximates the proposed initialization method. Furthermore, we suggest to adapt an efficient Monte Carlo approximation of $k$-means++ to AA++. In an extensive empirical evaluation of 13 real-world data sets of varying sizes and dimensionalities and considering two pre-processing strategies, we show that AA++ almost consistently outperforms all baselines, including the most frequently used ones.
comment: 20 pages, 13 figures, preprint
☆ Zero-shot-Learning Cross-Modality Data Translation Through Mutual Information Guided Stochastic Diffusion
Cross-modality data translation has attracted great interest in image computing. Deep generative models (\textit{e.g.}, GANs) show performance improvement in tackling those problems. Nevertheless, as a fundamental challenge in image translation, the problem of Zero-shot-Learning Cross-Modality Data Translation with fidelity remains unanswered. This paper proposes a new unsupervised zero-shot-learning method named Mutual Information guided Diffusion cross-modality data translation Model (MIDiffusion), which learns to translate the unseen source data to the target domain. The MIDiffusion leverages a score-matching-based generative model, which learns the prior knowledge in the target domain. We propose a differentiable local-wise-MI-Layer ($LMI$) for conditioning the iterative denoising sampling. The $LMI$ captures the identical cross-modality features in the statistical domain for the diffusion guidance; thus, our method does not require retraining when the source domain is changed, as it does not rely on any direct mapping between the source and target domains. This advantage is critical for applying cross-modality data translation methods in practice, as a reasonable amount of source domain dataset is not always available for supervised training. We empirically show the advanced performance of MIDiffusion in comparison with an influential group of generative models, including adversarial-based and other score-matching-based models.
☆ UPop: Unified and Progressive Pruning for Compressing Vision-Language Transformers
Real-world data contains a vast amount of multimodal information, among which vision and language are the two most representative modalities. Moreover, increasingly heavier models, e.g., Transformers, have attracted the attention of researchers to model compression. However, how to compress multimodal models, especially vison-language Transformers, is still under-explored. This paper proposes the \textbf{U}nified and \textbf{P}r\textbf{o}gressive \textbf{P}runing (UPop) as a universal vison-language Transformer compression framework, which incorporates 1) unifiedly searching multimodal subnets in a continuous optimization space from the original model, which enables automatic assignment of pruning ratios among compressible modalities and structures; 2) progressively searching and retraining the subnet, which maintains convergence between the search and retrain to attain higher compression ratios. Experiments on multiple generative and discriminative vision-language tasks, including Visual Reasoning, Image Caption, Visual Question Answer, Image-Text Retrieval, Text-Image Retrieval, and Image Classification, demonstrate the effectiveness and versatility of the proposed UPop framework.
comment: 16 pages, 5 figures, 13 tables
☆ Self-Consistent Velocity Matching of Probability Flows
We present a discretization-free scalable framework for solving a large class of mass-conserving partial differential equations (PDEs), including the time-dependent Fokker-Planck equation and the Wasserstein gradient flow. The main observation is that the time-varying velocity field of the PDE solution needs to be self-consistent: it must satisfy a fixed-point equation involving the flow characterized by the same velocity field. By parameterizing the flow as a time-dependent neural network, we propose an end-to-end iterative optimization framework called self-consistent velocity matching to solve this class of PDEs. Compared to existing approaches, our method does not suffer from temporal or spatial discretization, covers a wide range of PDEs, and scales to high dimensions. Experimentally, our method recovers analytical solutions accurately when they are available and achieves comparable or better performance in high dimensions with less training time compared to recent large-scale JKO-based methods that are designed for solving a more restrictive family of PDEs.
☆ Improving Monte Carlo Evaluation with Offline Data
Monte Carlo (MC) methods are the most widely used methods to estimate the performance of a policy. Given an interested policy, MC methods give estimates by repeatedly running this policy to collect samples and taking the average of the outcomes. Samples collected during this process are called online samples. To get an accurate estimate, MC methods consume massive online samples. When online samples are expensive, e.g., online recommendations and inventory management, we want to reduce the number of online samples while achieving the same estimate accuracy. To this end, we use off-policy MC methods that evaluate the interested policy by running a different policy called behavior policy. We design a tailored behavior policy such that the variance of the off-policy MC estimator is provably smaller than the ordinary MC estimator. Importantly, this tailored behavior policy can be efficiently learned from existing offline data, i,e., previously logged data, which are much cheaper than online samples. With reduced variance, our off-policy MC method requires fewer online samples to evaluate the performance of a policy compared with the ordinary MC method. Moreover, our off-policy MC estimator is always unbiased.
☆ A Bayesian Generative Adversarial Network (GAN) to Generate Synthetic Time-Series Data, Application in Combined Sewer Flow Prediction
Despite various breakthroughs in machine learning and data analysis techniques for improving smart operation and management of urban water infrastructures, some key limitations obstruct this progress. Among these shortcomings, the absence of freely available data due to data privacy or high costs of data gathering and the nonexistence of adequate rare or extreme events in the available data plays a crucial role. Here, Generative Adversarial Networks (GANs) can help overcome these challenges. In machine learning, generative models are a class of methods capable of learning data distribution to generate artificial data. In this study, we developed a GAN model to generate synthetic time series to balance our limited recorded time series data and improve the accuracy of a data-driven model for combined sewer flow prediction. We considered the sewer system of a small town in Germany as the test case. Precipitation and inflow to the storage tanks are used for the Data-Driven model development. The aim is to predict the flow using precipitation data and examine the impact of data augmentation using synthetic data in model performance. Results show that GAN can successfully generate synthetic time series from real data distribution, which helps more accurate peak flow prediction. However, the model without data augmentation works better for dry weather prediction. Therefore, an ensemble model is suggested to combine the advantages of both models.
comment: Accepted in WDSA/CCWI 2022 Conference
☆ Preserving local densities in low-dimensional embeddings
Low-dimensional embeddings and visualizations are an indispensable tool for analysis of high-dimensional data. State-of-the-art methods, such as tSNE and UMAP, excel in unveiling local structures hidden in high-dimensional data and are therefore routinely applied in standard analysis pipelines in biology. We show, however, that these methods fail to reconstruct local properties, such as relative differences in densities (Fig. 1) and that apparent differences in cluster size can arise from computational artifact caused by differing sample sizes (Fig. 2). Providing a theoretical analysis of this issue, we then suggest dtSNE, which approximately conserves local densities. In an extensive study on synthetic benchmark and real world data comparing against five state-of-the-art methods, we empirically show that dtSNE provides similar global reconstruction, but yields much more accurate depictions of local distances and relative densities.
☆ A relaxed proximal gradient descent algorithm for convergent plug-and-play with proximal denoiser
This paper presents a new convergent Plug-and-Play (PnP) algorithm. PnP methods are efficient iterative algorithms for solving image inverse problems formulated as the minimization of the sum of a data-fidelity term and a regularization term. PnP methods perform regularization by plugging a pre-trained denoiser in a proximal algorithm, such as Proximal Gradient Descent (PGD). To ensure convergence of PnP schemes, many works study specific parametrizations of deep denoisers. However, existing results require either unverifiable or suboptimal hypotheses on the denoiser, or assume restrictive conditions on the parameters of the inverse problem. Observing that these limitations can be due to the proximal algorithm in use, we study a relaxed version of the PGD algorithm for minimizing the sum of a convex function and a weakly convex one. When plugged with a relaxed proximal denoiser, we show that the proposed PnP-$\alpha$PGD algorithm converges for a wider range of regularization parameters, thus allowing more accurate image restoration.
☆ Convolutional autoencoder for the spatiotemporal latent representation of turbulence
Turbulence is characterised by chaotic dynamics and a high-dimensional state space, which make the phenomenon challenging to predict. However, turbulent flows are often characterised by coherent spatiotemporal structures, such as vortices or large-scale modes, which can help obtain a latent description of turbulent flows. However, current approaches are often limited by either the need to use some form of thresholding on quantities defining the isosurfaces to which the flow structures are associated or the linearity of traditional modal flow decomposition approaches, such as those based on proper orthogonal decomposition. This problem is exacerbated in flows that exhibit extreme events, which are rare and sudden changes in a turbulent state. The goal of this paper is to obtain an efficient and accurate reduced-order latent representation of a turbulent flow that exhibits extreme events. Specifically, we employ a three-dimensional multiscale convolutional autoencoder (CAE) to obtain such latent representation. We apply it to a three-dimensional turbulent flow. We show that the Multiscale CAE is efficient, requiring less than 10% degrees of freedom than proper orthogonal decomposition for compressing the data and is able to accurately reconstruct flow states related to extreme events. The proposed deep learning architecture opens opportunities for nonlinear reduced-order modeling of turbulent flows from data.
☆ The passive symmetries of machine learning
Any representation of data involves arbitrary investigator choices. Because those choices are external to the data-generating process, each choice leads to an exact symmetry, corresponding to the group of transformations that takes one possible representation to another. These are the passive symmetries; they include coordinate freedom, gauge symmetry and units covariance, all of which have led to important results in physics. Our goal is to understand the implications of passive symmetries for machine learning: Which passive symmetries play a role (e.g., permutation symmetry in graph neural networks)? What are dos and don'ts in machine learning practice? We assay conditions under which passive symmetries can be implemented as group equivariances. We also discuss links to causal modeling, and argue that the implementation of passive symmetries is particularly valuable when the goal of the learning problem is to generalize out of sample. While this paper is purely conceptual, we believe that it can have a significant impact on helping machine learning make the transition that took place for modern physics in the first half of the Twentieth century.
☆ On the Initialisation of Wide Low-Rank Feedforward Neural Networks
The edge-of-chaos dynamics of wide randomly initialized low-rank feedforward networks are analyzed. Formulae for the optimal weight and bias variances are extended from the full-rank to low-rank setting and are shown to follow from multiplicative scaling. The principle second order effect, the variance of the input-output Jacobian, is derived and shown to increase as the rank to width ratio decreases. These results inform practitioners how to randomly initialize feedforward networks with a reduced number of learnable parameters while in the same ambient dimension, allowing reductions in the computational cost and memory constraints of the associated network.
☆ Dissecting the Effects of SGD Noise in Distinct Regimes of Deep Learning
Understanding when the noise in stochastic gradient descent (SGD) affects generalization of deep neural networks remains a challenge, complicated by the fact that networks can operate in distinct training regimes. Here we study how the magnitude of this noise $T$ affects performance as the size of the training set $P$ and the scale of initialization $\alpha$ are varied. For gradient descent, $\alpha$ is a key parameter that controls if the network is `lazy' ($\alpha\gg 1$) or instead learns features ($\alpha\ll 1$). For classification of MNIST and CIFAR10 images, our central results are: (i) obtaining phase diagrams for performance in the $(\alpha,T)$ plane. They show that SGD noise can be detrimental or instead useful depending on the training regime. Moreover, although increasing $T$ or decreasing $\alpha$ both allow the net to escape the lazy regime, these changes can have opposite effects on performance. (ii) Most importantly, we find that key dynamical quantities (including the total variations of weights during training) depend on both $T$ and $P$ as power laws, and the characteristic temperature $T_c$, where the noise of SGD starts affecting performance, is a power law of $P$. These observations indicate that a key effect of SGD noise occurs late in training, by affecting the stopping process whereby all data are fitted. We argue that due to SGD noise, nets must develop a stronger `signal', i.e. larger informative weights, to fit the data, leading to a longer training time. The same effect occurs at larger training set $P$. We confirm this view in the perceptron model, where signal and noise can be precisely measured. Interestingly, exponents characterizing the effect of SGD depend on the density of data near the decision boundary, as we explain.
comment: 18 pages, 14 figures
☆ Are Defenses for Graph Neural Networks Robust? NeurIPS 2022
A cursory reading of the literature suggests that we have made a lot of progress in designing effective adversarial defenses for Graph Neural Networks (GNNs). Yet, the standard methodology has a serious flaw - virtually all of the defenses are evaluated against non-adaptive attacks leading to overly optimistic robustness estimates. We perform a thorough robustness analysis of 7 of the most popular defenses spanning the entire spectrum of strategies, i.e., aimed at improving the graph, the architecture, or the training. The results are sobering - most defenses show no or only marginal improvement compared to an undefended baseline. We advocate using custom adaptive attacks as a gold standard and we outline the lessons we learned from successfully designing such attacks. Moreover, our diverse collection of perturbed graphs forms a (black-box) unit test offering a first glance at a model's robustness.
comment: 34 pages, 36th Conference on Neural Information Processing Systems (NeurIPS 2022)
☆ Time Series Forecasting via Semi-Asymmetric Convolutional Architecture with Global Atrous Sliding Window
The proposed method in this paper is designed to address the problem of time series forecasting. Although some exquisitely designed models achieve excellent prediction performances, how to extract more useful information and make accurate predictions is still an open issue. Most of modern models only focus on a short range of information, which are fatal for problems such as time series forecasting which needs to capture long-term information characteristics. As a result, the main concern of this work is to further mine relationship between local and global information contained in time series to produce more precise predictions. In this paper, to satisfactorily realize the purpose, we make three main contributions that are experimentally verified to have performance advantages. Firstly, original time series is transformed into difference sequence which serves as input to the proposed model. And secondly, we introduce the global atrous sliding window into the forecasting model which references the concept of fuzzy time series to associate relevant global information with temporal data within a time period and utilizes central-bidirectional atrous algorithm to capture underlying-related features to ensure validity and consistency of captured data. Thirdly, a variation of widely-used asymmetric convolution which is called semi-asymmetric convolution is devised to more flexibly extract relationships in adjacent elements and corresponding associated global features with adjustable ranges of convolution on vertical and horizontal directions. The proposed model in this paper achieves state-of-the-art on most of time series datasets provided compared with competitive modern models.
comment: 13pages,8 figures
☆ The Flan Collection: Designing Data and Methods for Effective Instruction Tuning
We study the design decisions of publicly available instruction tuning methods, and break down the development of Flan 2022 (Chung et al., 2022). Through careful ablation studies on the Flan Collection of tasks and methods, we tease apart the effect of design decisions which enable Flan-T5 to outperform prior work by 3-17%+ across evaluation settings. We find task balancing and enrichment techniques are overlooked but critical to effective instruction tuning, and in particular, training with mixed prompt settings (zero-shot, few-shot, and chain-of-thought) actually yields stronger (2%+) performance in all settings. In further experiments, we show Flan-T5 requires less finetuning to converge higher and faster than T5 on single downstream tasks, motivating instruction-tuned models as more computationally-efficient starting checkpoints for new tasks. Finally, to accelerate research on instruction tuning, we make the Flan 2022 collection of datasets, templates, and methods publicly available at https://github.com/google-research/FLAN/tree/main/flan/v2.
☆ Improved distinct bone segmentation in upper-body CT through multi-resolution networks
Purpose: Automated distinct bone segmentation from CT scans is widely used in planning and navigation workflows. U-Net variants are known to provide excellent results in supervised semantic segmentation. However, in distinct bone segmentation from upper body CTs a large field of view and a computationally taxing 3D architecture are required. This leads to low-resolution results lacking detail or localisation errors due to missing spatial context when using high-resolution inputs. Methods: We propose to solve this problem by using end-to-end trainable segmentation networks that combine several 3D U-Nets working at different resolutions. Our approach, which extends and generalizes HookNet and MRN, captures spatial information at a lower resolution and skips the encoded information to the target network, which operates on smaller high-resolution inputs. We evaluated our proposed architecture against single resolution networks and performed an ablation study on information concatenation and the number of context networks. Results: Our proposed best network achieves a median DSC of 0.86 taken over all 125 segmented bone classes and reduces the confusion among similar-looking bones in different locations. These results outperform our previously published 3D U-Net baseline results on the task and distinct-bone segmentation results reported by other groups. Conclusion: The presented multi-resolution 3D U-Nets address current shortcomings in bone segmentation from upper-body CT scans by allowing for capturing a larger field of view while avoiding the cubic growth of the input pixels and intermediate computations that quickly outgrow the computational capacities in 3D. The approach thus improves the accuracy and efficiency of distinct bone segmentation from upper-body CT.
comment: Under submission
☆ Enhancing Hyper-To-Real Space Projections Through Euclidean Norm Meta-Heuristic Optimization
The continuous computational power growth in the last decades has made solving several optimization problems significant to humankind a tractable task; however, tackling some of them remains a challenge due to the overwhelming amount of candidate solutions to be evaluated, even by using sophisticated algorithms. In such a context, a set of nature-inspired stochastic methods, called meta-heuristic optimization, can provide robust approximate solutions to different kinds of problems with a small computational burden, such as derivative-free real function optimization. Nevertheless, these methods may converge to inadequate solutions if the function landscape is too harsh, e.g., enclosing too many local optima. Previous works addressed this issue by employing a hypercomplex representation of the search space, like quaternions, where the landscape becomes smoother and supposedly easier to optimize. Under this approach, meta-heuristic computations happen in the hypercomplex space, whereas variables are mapped back to the real domain before function evaluation. Despite this latter operation being performed by the Euclidean norm, we have found that after the optimization procedure has finished, it is usually possible to obtain even better solutions by employing the Minkowski $p$-norm instead and fine-tuning $p$ through an auxiliary sub-problem with neglecting additional cost and no hyperparameters. Such behavior was observed in eight well-established benchmarking functions, thus fostering a new research direction for hypercomplex meta-heuristic optimization.
☆ Reinforcement learning and decision making via single-photon quantum walks
Variational quantum algorithms represent a promising approach to quantum machine learning where classical neural networks are replaced by parametrized quantum circuits. Here, we present a variational approach to quantize projective simulation (PS), a reinforcement learning model aimed at interpretable artificial intelligence. Decision making in PS is modeled as a random walk on a graph describing the agent's memory. To implement the quantized model, we consider quantum walks of single photons in a lattice of tunable Mach-Zehnder interferometers. We propose variational algorithms tailored to reinforcement learning tasks, and we show, using an example from transfer learning, that the quantized PS learning model can outperform its classical counterpart. Finally, we discuss the role of quantum interference for training and decision making, paving the way for realizations of interpretable quantum learning agents.
comment: 10+6 pages, 6+5 figures, 2 tables. F. Flamini and M. Krumm contributed equally to this work
☆ Automated Sentiment and Hate Speech Analysis of Facebook Data by Employing Multilingual Transformer Models
In recent years, there has been a heightened consensus within academia and in the public discourse that Social Media Platforms (SMPs), amplify the spread of hateful and negative sentiment content. Researchers have identified how hateful content, political propaganda, and targeted messaging contributed to real-world harms including insurrections against democratically elected governments, genocide, and breakdown of social cohesion due to heightened negative discourse towards certain communities in parts of the world. To counter these issues, SMPs have created semi-automated systems that can help identify toxic speech. In this paper we analyse the statistical distribution of hateful and negative sentiment contents within a representative Facebook dataset (n= 604,703) scrapped through 648 public Facebook pages which identify themselves as proponents (and followers) of far-right Hindutva actors. These pages were identified manually using keyword searches on Facebook and on CrowdTangleand classified as far-right Hindutva pages based on page names, page descriptions, and discourses shared on these pages. We employ state-of-the-art, open-source XLM-T multilingual transformer-based language models to perform sentiment and hate speech analysis of the textual contents shared on these pages over a period of 5.5 years. The result shows the statistical distributions of the predicted sentiment and the hate speech labels; top actors, and top page categories. We further discuss the benchmark performances and limitations of these pre-trained language models.
☆ Collision-aware In-hand 6D Object Pose Estimation using Multiple Vision-based Tactile Sensors ICRA
In this paper, we address the problem of estimating the in-hand 6D pose of an object in contact with multiple vision-based tactile sensors. We reason on the possible spatial configurations of the sensors along the object surface. Specifically, we filter contact hypotheses using geometric reasoning and a Convolutional Neural Network (CNN), trained on simulated object-agnostic images, to promote those that better comply with the actual tactile images from the sensors. We use the selected sensors configurations to optimize over the space of 6D poses using a Gradient Descent-based approach. We finally rank the obtained poses by penalizing those that are in collision with the sensors. We carry out experiments in simulation using the DIGIT vision-based sensor with several objects, from the standard YCB model set. The results demonstrate that our approach estimates object poses that are compatible with actual object-sensor contacts in $87.5\%$ of cases while reaching an average positional error in the order of $2$ centimeters. Our analysis also includes qualitative results of experiments with a real DIGIT sensor.
comment: Accepted for publication at 2023 IEEE International Conference on Robotics and Automation (ICRA)
☆ Exploring QSAR Models for Activity-Cliff Prediction
Pairs of similar compounds that only differ by a small structural modification but exhibit a large difference in their binding affinity for a given target are known as activity cliffs (ACs). It has been hypothesised that quantitative structure-activity relationship (QSAR) models struggle to predict ACs and that ACs thus form a major source of prediction error. However, a study to explore the AC-prediction power of modern QSAR methods and its relationship to general QSAR-prediction performance is lacking. We systematically construct nine distinct QSAR models by combining three molecular representation methods (extended-connectivity fingerprints, physicochemical-descriptor vectors and graph isomorphism networks) with three regression techniques (random forests, k-nearest neighbours and multilayer perceptrons); we then use each resulting model to classify pairs of similar compounds as ACs or non-ACs and to predict the activities of individual molecules in three case studies: dopamine receptor D2, factor Xa, and SARS-CoV-2 main protease. We observe low AC-sensitivity amongst the tested models when the activities of both compounds are unknown, but a substantial increase in AC-sensitivity when the actual activity of one of the compounds is given. Graph isomorphism features are found to be competitive with or superior to classical molecular representations for AC-classification and can thus be employed as baseline AC-prediction models or simple compound-optimisation tools. For general QSAR-prediction, however, extended-connectivity fingerprints still consistently deliver the best performance. Our results provide strong support for the hypothesis that indeed QSAR methods frequently fail to predict ACs. We propose twin-network training for deep learning models as a potential future pathway to increase AC-sensitivity and thus overall QSAR performance.
comment: Submitted to Journal of Cheminformatics
☆ An Efficient Solution to s-Rectangular Robust Markov Decision Processes
We present an efficient robust value iteration for \texttt{s}-rectangular robust Markov Decision Processes (MDPs) with a time complexity comparable to standard (non-robust) MDPs which is significantly faster than any existing method. We do so by deriving the optimal robust Bellman operator in concrete forms using our $L_p$ water filling lemma. We unveil the exact form of the optimal policies, which turn out to be novel threshold policies with the probability of playing an action proportional to its advantage.
comment: arXiv admin note: substantial text overlap with arXiv:2205.14327
☆ Tricking AI chips into Simulating the Human Brain: A Detailed Performance Analysis
Challenging the Nvidia monopoly, dedicated AI-accelerator chips have begun emerging for tackling the computational challenge that the inference and, especially, the training of modern deep neural networks (DNNs) poses to modern computers. The field has been ridden with studies assessing the performance of these contestants across various DNN model types. However, AI-experts are aware of the limitations of current DNNs and have been working towards the fourth AI wave which will, arguably, rely on more biologically inspired models, predominantly on spiking neural networks (SNNs). At the same time, GPUs have been heavily used for simulating such models in the field of computational neuroscience, yet AI-chips have not been tested on such workloads. The current paper aims at filling this important gap by evaluating multiple, cutting-edge AI-chips (Graphcore IPU, GroqChip, Nvidia GPU with Tensor Cores and Google TPU) on simulating a highly biologically detailed model of a brain region, the inferior olive (IO). This IO application stress-tests the different AI-platforms for highlighting architectural tradeoffs by varying its compute density, memory requirements and floating-point numerical accuracy. Our performance analysis reveals that the simulation problem maps extremely well onto the GPU and TPU architectures, which for networks of 125,000 cells leads to a 28x respectively 1,208x speedup over CPU runtimes. At this speed, the TPU sets a new record for largest real-time IO simulation. The GroqChip outperforms both platforms for small networks but, due to implementing some floating-point operations at reduced accuracy, is found not yet usable for brain simulation.
comment: 11 pages, 4 figures
☆ Transport with Support: Data-Conditional Diffusion Bridges
The dynamic Schr\"odinger bridge problem provides an appealing setting for solving optimal transport problems by learning non-linear diffusion processes using efficient iterative solvers. Recent works have demonstrated state-of-the-art results (eg. in modelling single-cell embryo RNA sequences or sampling from complex posteriors) but are limited to learning bridges with only initial and terminal constraints. Our work extends this paradigm by proposing the Iterative Smoothing Bridge (ISB). We integrate Bayesian filtering and optimal control into learning the diffusion process, enabling constrained stochastic processes governed by sparse observations at intermediate stages and terminal constraints. We assess the effectiveness of our method on synthetic and real-world data and show that the ISB generalises well to high-dimensional data, is computationally efficient, and provides accurate estimates of the marginals at intermediate and terminal times.
comment: 23 pages, 11 figures
☆ Active Learning-based Domain Adaptive Localized Polynomial Chaos Expansion
The paper presents a novel methodology to build surrogate models of complicated functions by an active learning-based sequential decomposition of the input random space and construction of localized polynomial chaos expansions, referred to as domain adaptive localized polynomial chaos expansion (DAL-PCE). The approach utilizes sequential decomposition of the input random space into smaller sub-domains approximated by low-order polynomial expansions. This allows approximation of functions with strong nonlinearties, discontinuities, and/or singularities. Decomposition of the input random space and local approximations alleviates the Gibbs phenomenon for these types of problems and confines error to a very small vicinity near the non-linearity. The global behavior of the surrogate model is therefore significantly better than existing methods as shown in numerical examples. The whole process is driven by an active learning routine that uses the recently proposed $\Theta$ criterion to assess local variance contributions. The proposed approach balances both \emph{exploitation} of the surrogate model and \emph{exploration} of the input random space and thus leads to efficient and accurate approximation of the original mathematical model. The numerical results show the superiority of the DAL-PCE in comparison to (i) a single global polynomial chaos expansion and (ii) the recently proposed stochastic spectral embedding (SSE) method developed as an accurate surrogate model and which is based on a similar domain decomposition process. This method represents general framework upon which further extensions and refinements can be based, and which can be combined with any technique for non-intrusive polynomial chaos expansion construction.
☆ DiffSTG: Probabilistic Spatio-Temporal Graph Forecasting with Denoising Diffusion Models
Spatio-temporal graph neural networks (STGNN) have emerged as the dominant model for spatio-temporal graph (STG) forecasting. Despite their success, they fail to model intrinsic uncertainties within STG data, which cripples their practicality in downstream tasks for decision-making. To this end, this paper focuses on probabilistic STG forecasting, which is challenging due to the difficulty in modeling uncertainties and complex ST dependencies. In this study, we present the first attempt to generalize the popular denoising diffusion probabilistic models to STGs, leading to a novel non-autoregressive framework called DiffSTG, along with the first denoising network UGnet for STG in the framework. Our approach combines the spatio-temporal learning capabilities of STGNNs with the uncertainty measurements of diffusion models. Extensive experiments validate that DiffSTG reduces the Continuous Ranked Probability Score (CRPS) by 4%-14%, and Root Mean Squared Error (RMSE) by 2%-7% over existing methods on three real-world datasets.
☆ Learning Data Representations with Joint Diffusion Models
We introduce a joint diffusion model that simultaneously learns meaningful internal representations fit for both generative and predictive tasks. Joint machine learning models that allow synthesizing and classifying data often offer uneven performance between those tasks or are unstable to train. In this work, we depart from a set of empirical observations that indicate the usefulness of internal representations built by contemporary deep diffusion-based generative models in both generative and predictive settings. We then introduce an extension of the vanilla diffusion model with a classifier that allows for stable joint training with shared parametrization between those objectives. The resulting joint diffusion model offers superior performance across various tasks, including generative modeling, semi-supervised classification, and domain adaptation.
☆ Scheduling Inference Workloads on Distributed Edge Clusters with Reinforcement Learning
Many real-time applications (e.g., Augmented/Virtual Reality, cognitive assistance) rely on Deep Neural Networks (DNNs) to process inference tasks. Edge computing is considered a key infrastructure to deploy such applications, as moving computation close to the data sources enables us to meet stringent latency and throughput requirements. However, the constrained nature of edge networks poses several additional challenges to the management of inference workloads: edge clusters can not provide unlimited processing power to DNN models, and often a trade-off between network and processing time should be considered when it comes to end-to-end delay requirements. In this paper, we focus on the problem of scheduling inference queries on DNN models in edge networks at short timescales (i.e., few milliseconds). By means of simulations, we analyze several policies in the realistic network settings and workloads of a large ISP, highlighting the need for a dynamic scheduling policy that can adapt to network conditions and workloads. We therefore design ASET, a Reinforcement Learning based scheduling algorithm able to adapt its decisions according to the system conditions. Our results show that ASET effectively provides the best performance compared to static policies when scheduling over a distributed pool of edge resources.
☆ Anti-Exploration by Random Network Distillation
Despite the success of Random Network Distillation (RND) in various domains, it was shown as not discriminative enough to be used as an uncertainty estimator for penalizing out-of-distribution actions in offline reinforcement learning. In this paper, we revisit these results and show that, with a naive choice of conditioning for the RND prior, it becomes infeasible for the actor to effectively minimize the anti-exploration bonus and discriminativity is not an issue. We show that this limitation can be avoided with conditioning based on Feature-wise Linear Modulation (FiLM), resulting in a simple and efficient ensemble-free algorithm based on Soft Actor-Critic. We evaluate it on the D4RL benchmark, showing that it is capable of achieving performance comparable to ensemble-based methods and outperforming ensemble-free approaches by a wide margin.
comment: Source code: https://github.com/tinkoff-ai/sac-rnd
☆ Policy Gradient for s-Rectangular Robust Markov Decision Processes
We present a novel robust policy gradient method (RPG) for s-rectangular robust Markov Decision Processes (MDPs). We are the first to derive the adversarial kernel in a closed form and demonstrate that it is a one-rank perturbation of the nominal kernel. This allows us to derive an RPG that is similar to the one used in non-robust MDPs, except with a robust Q-value function and an additional correction term. Both robust Q-values and correction terms are efficiently computable, thus the time complexity of our method matches that of non-robust MDPs, which is significantly faster compared to existing black box methods.
☆ Support Exploration Algorithm for Sparse Support Recovery
We introduce a new algorithm promoting sparsity called {\it Support Exploration Algorithm (SEA)} and analyze it in the context of support recovery/model selection problems.The algorithm can be interpreted as an instance of the {\it straight-through estimator (STE)} applied to the resolution of a sparse linear inverse problem. SEA uses a non-sparse exploratory vector and makes it evolve in the input space to select the sparse support. We put to evidence an oracle update rule for the exploratory vector and consider the STE update. The theoretical analysis establishes general sufficient conditions of support recovery. The general conditions are specialized to the case where the matrix $A$ performing the linear measurements satisfies the {\it Restricted Isometry Property (RIP)}.Experiments show that SEA can efficiently improve the results of any algorithm. Because of its exploratory nature, SEA also performs remarkably well when the columns of $A$ are strongly coherent.
☆ Sport Task: Fine Grained Action Detection and Classification of Table Tennis Strokes from Videos for MediaEval 2022
Sports video analysis is a widespread research topic. Its applications are very diverse, like events detection during a match, video summary, or fine-grained movement analysis of athletes. As part of the MediaEval 2022 benchmarking initiative, this task aims at detecting and classifying subtle movements from sport videos. We focus on recordings of table tennis matches. Conducted since 2019, this task provides a classification challenge from untrimmed videos recorded under natural conditions with known temporal boundaries for each stroke. Since 2021, the task also provides a stroke detection challenge from unannotated, untrimmed videos. This year, the training, validation, and test sets are enhanced to ensure that all strokes are represented in each dataset. The dataset is now similar to the one used in [1, 2]. This research is intended to build tools for coaches and athletes who want to further evaluate their sport performances.
comment: MediaEval 2022 Workshop, Jan 2023, Bergen, Norway. arXiv admin note: substantial text overlap with arXiv:2112.11384
☆ Skill Decision Transformer
Recent work has shown that Large Language Models (LLMs) can be incredibly effective for offline reinforcement learning (RL) by representing the traditional RL problem as a sequence modelling problem (Chen et al., 2021; Janner et al., 2021). However many of these methods only optimize for high returns, and may not extract much information from a diverse dataset of trajectories. Generalized Decision Transformers (GDTs) (Furuta et al., 2021) have shown that utilizing future trajectory information, in the form of information statistics, can help extract more information from offline trajectory data. Building upon this, we propose Skill Decision Transformer (Skill DT). Skill DT draws inspiration from hindsight relabelling (Andrychowicz et al., 2017) and skill discovery methods to discover a diverse set of primitive behaviors, or skills. We show that Skill DT can not only perform offline state-marginal matching (SMM), but can discovery descriptive behaviors that can be easily sampled. Furthermore, we show that through purely reward-free optimization, Skill DT is still competitive with supervised offline RL approaches on the D4RL benchmark. The code and videos can be found on our project page: https://github.com/shyamsn97/skill-dt
☆ BALANCE: Bayesian Linear Attribution for Root Cause Localization SIGMOD 2023
Root Cause Analysis (RCA) plays an indispensable role in distributed data system maintenance and operations, as it bridges the gap between fault detection and system recovery. Existing works mainly study multidimensional localization or graph-based root cause localization. This paper opens up the possibilities of exploiting the recently developed framework of explainable AI (XAI) for the purpose of RCA. In particular, we propose BALANCE (BAyesian Linear AttributioN for root CausE localization), which formulates the problem of RCA through the lens of attribution in XAI and seeks to explain the anomalies in the target KPIs by the behavior of the candidate root causes. BALANCE consists of three innovative components. First, we propose a Bayesian multicollinear feature selection (BMFS) model to predict the target KPIs given the candidate root causes in a forward manner while promoting sparsity and concurrently paying attention to the correlation between the candidate root causes. Second, we introduce attribution analysis to compute the attribution score for each candidate in a backward manner. Third, we merge the estimated root causes related to each KPI if there are multiple KPIs. We extensively evaluate the proposed BALANCE method on one synthesis dataset as well as three real-world RCA tasks, that is, bad SQL localization, container fault localization, and fault type diagnosis for Exathlon. Results show that BALANCE outperforms the state-of-the-art (SOTA) methods in terms of accuracy with the least amount of running time, and achieves at least $6\%$ notably higher accuracy than SOTA methods for real tasks. BALANCE has been deployed to production to tackle real-world RCA problems, and the online results further advocate its usage for real-time diagnosis in distributed data systems.
comment: Accepted by SIGMOD 2023; 15 pages
☆ NP-Match: Towards a New Probabilistic Model for Semi-Supervised Learning ICML 2022
Semi-supervised learning (SSL) has been widely explored in recent years, and it is an effective way of leveraging unlabeled data to reduce the reliance on labeled data. In this work, we adjust neural processes (NPs) to the semi-supervised image classification task, resulting in a new method named NP-Match. NP-Match is suited to this task for two reasons. Firstly, NP-Match implicitly compares data points when making predictions, and as a result, the prediction of each unlabeled data point is affected by the labeled data points that are similar to it, which improves the quality of pseudo-labels. Secondly, NP-Match is able to estimate uncertainty that can be used as a tool for selecting unlabeled samples with reliable pseudo-labels. Compared with uncertainty-based SSL methods implemented with Monte-Carlo (MC) dropout, NP-Match estimates uncertainty with much less computational overhead, which can save time at both the training and the testing phases. We conducted extensive experiments on five public datasets under three semi-supervised image classification settings, namely, the standard semi-supervised image classification, the imbalanced semi-supervised image classification, and the multi-label semi-supervised image classification, and NP-Match outperforms state-of-the-art (SOTA) approaches or achieves competitive results on them, which shows the effectiveness of NP-Match and its potential for SSL. The codes are at https://github.com/Jianf-Wang/NP-Match
comment: An journal version of our previous ICML 2022 paper arXiv:2207.01066 . Codes are available at: https://github.com/Jianf-Wang/NP-Match
☆ Learning Against Distributional Uncertainty: On the Trade-off Between Robustness and Specificity
Trustworthy machine learning aims at combating distributional uncertainties in training data distributions compared to population distributions. Typical treatment frameworks include the Bayesian approach, (min-max) distributionally robust optimization (DRO), and regularization. However, two issues have to be raised: 1) All these methods are biased estimators of the true optimal cost; 2) the prior distribution in the Bayesian method, the radius of the distributional ball in the DRO method, and the regularizer in the regularization method are difficult to specify. This paper studies a new framework that unifies the three approaches and that addresses the two challenges mentioned above. The asymptotic properties (e.g., consistency and asymptotic normalities), non-asymptotic properties (e.g., unbiasedness and generalization error bound), and a Monte--Carlo-based solution method of the proposed model are studied. The new model reveals the trade-off between the robustness to the unseen data and the specificity to the training data.
comment: 23 Pages, 3 Figures
Review of methods for automatic cerebral microbleeds detection
Cerebral microbleeds detection is an important and challenging task. With the gaining popularity of the MRI, the ability to detect cerebral microbleeds also raises. Unfortunately, for radiologists, it is a time-consuming and laborious procedure. For this reason, various solutions to automate this process have been proposed for several years, but none of them is currently used in medical practice. In this context, the need to systematize the existing knowledge and best practices has been recognized as a factor facilitating the imminent synthesis of a real CMBs detection system practically applicable in medicine. To the best of our knowledge, all available publications regarding automatic cerebral microbleeds detection have been gathered, described, and assessed in this paper in order to distinguish the current research state and provide a starting point for future studies.
comment: 32 pages, 6 figures, 3 tables, 174 references
☆ Holistic Graph-based Motion Prediction ICRA 2023
Motion prediction for automated vehicles in complex environments is a difficult task that is to be mastered when automated vehicles are to be used in arbitrary situations. Many factors influence the future motion of traffic participants starting with traffic rules and reaching from the interaction between each other to personal habits of human drivers. Therefore we present a novel approach for a graph-based prediction based on a heterogeneous holistic graph representation that combines temporal information, properties and relations between traffic participants as well as relations with static elements like the road network. The information are encoded through different types of nodes and edges that both are enriched with arbitrary features. We evaluated the approach on the INTERACTION and the Argoverse dataset and conducted an informative ablation study to demonstrate the benefit of different types of information for the motion prediction quality.
comment: Accepted on ICRA 2023
☆ Low Complexity Adaptive Machine Learning Approaches for End-to-End Latency Prediction
Software Defined Networks have opened the door to statistical and AI-based techniques to improve efficiency of networking. Especially to ensure a certain Quality of Service (QoS) for specific applications by routing packets with awareness on content nature (VoIP, video, files, etc.) and its needs (latency, bandwidth, etc.) to use efficiently resources of a network. Monitoring and predicting various Key Performance Indicators (KPIs) at any level may handle such problems while preserving network bandwidth. The question addressed in this work is the design of efficient, low-cost adaptive algorithms for KPI estimation, monitoring and prediction. We focus on end-to-end latency prediction, for which we illustrate our approaches and results on data obtained from a public generator provided after the recent international challenge on GNN [12]. In this paper, we improve our previously proposed low-cost estimators [6] by adding the adaptive dimension, and show that the performances are minimally modified while gaining the ability to track varying networks.
☆ Population-wise Labeling of Sulcal Graphs using Multi-graph Matching
Population-wise matching of the cortical fold is necessary to identify biomarkers of neurological or psychiatric disorders. The difficulty comes from the massive interindividual variations in the morphology and spatial organization of the folds. This task is challenging at both methodological and conceptual levels. In the widely used registration-based techniques, these variations are considered as noise and the matching of folds is only implicit. Alternative approaches are based on the extraction and explicit identification of the cortical folds. In particular, representing cortical folding patterns as graphs of sulcal basins-termed sulcal graphs-enables to formalize the task as a graph-matching problem. In this paper, we propose to address the problem of sulcal graph matching directly at the population level using multi-graph matching techniques. First, we motivate the relevance of multi-graph matching framework in this context. We then introduce a procedure to generate populations of artificial sulcal graphs, which allows us benchmarking several state of the art multi-graph matching methods. Our results on both artificial and real data demonstrate the effectiveness of multi-graph matching techniques to obtain a population-wise consistent labeling of cortical folds at the sulcal basins level.
☆ Domain-Generalizable Multiple-Domain Clustering
Accurately clustering high-dimensional measurements is vital for adequately analyzing scientific data. Deep learning machinery has remarkably improved clustering capabilities in recent years due to its ability to extract meaningful representations. In this work, we are given unlabeled samples from multiple source domains, and we aim to learn a shared classifier that assigns the examples to various clusters. Evaluation is done by using the classifier for predicting cluster assignments in a previously unseen domain. This setting generalizes the problem of unsupervised domain generalization to the case in which no supervised learning samples are given (completely unsupervised). Towards this goal, we present an end-to-end model and evaluate its capabilities on several multi-domain image datasets. Specifically, we demonstrate that our model is more accurate than schemes that require fine-tuning using samples from the target domain or some level of supervision.
comment: 12 pages, 5 figures
☆ Real-Time Outlier Detection with Dynamic Process Limits
Anomaly detection methods are part of the systems where rare events may endanger an operation's profitability, safety, and environmental aspects. Although many state-of-the-art anomaly detection methods were developed to date, their deployment is limited to the operation conditions present during the model training. Online anomaly detection brings the capability to adapt to data drifts and change points that may not be represented during model development resulting in prolonged service life. This paper proposes an online anomaly detection algorithm for existing real-time infrastructures where low-latency detection is required and novel patterns in data occur unpredictably. The online inverse cumulative distribution-based approach is introduced to eliminate common problems of offline anomaly detectors, meanwhile providing dynamic process limits to normal operation. The benefit of the proposed method is the ease of use, fast computation, and deployability as shown in two case studies of real microgrid operation data.
comment: 7 pages, 4 figures, 24th International Conference on Process Control
☆ Quantum contextual bandits and recommender systems for quantum data
We study a recommender system for quantum data using the linear contextual bandit framework. In each round, a learner receives an observable (the context) and has to recommend from a finite set of unknown quantum states (the actions) which one to measure. The learner has the goal of maximizing the reward in each round, that is the outcome of the measurement on the unknown state. Using this model we formulate the low energy quantum state recommendation problem where the context is a Hamiltonian and the goal is to recommend the state with the lowest energy. For this task, we study two families of contexts: the Ising model and a generalized cluster model. We observe that if we interpret the actions as different phases of the models then the recommendation is done by classifying the correct phase of the given Hamiltonian and the strategy can be interpreted as an online quantum phase classifier.
comment: 15 pages, 9 figures
☆ Recurrences reveal shared causal drivers of complex time series
Many experimental time series measurements share an unobserved causal driver. Examples include genes targeted by transcription factors, ocean flows influenced by large-scale atmospheric currents, and motor circuits steered by descending neurons. Reliably inferring this unseen driving force is necessary to understand the intermittent nature of top-down control schemes in diverse biological and engineered systems. Here, we introduce a new unsupervised learning algorithm that uses recurrences in time series measurements to gradually reconstruct an unobserved driving signal. Drawing on the mathematical theory of skew-product dynamical systems, we identify recurrence events shared across response time series, which implicitly define a recurrence graph with glass-like structure. As the amount or quality of observed data improves, this recurrence graph undergoes a percolation transition manifesting as weak ergodicity breaking for random walks on the induced landscape -- revealing the shared driver's dynamics, even in the presence of strongly corrupted or noisy measurements. Across several thousand random dynamical systems, we empirically quantify the dependence of reconstruction accuracy on the rate of information transfer from a chaotic driver to the response systems, and we find that effective reconstruction proceeds through gradual approximation of the driver's dominant unstable periodic orbits. Through extensive benchmarks against classical and neural-network-based signal processing techniques, we demonstrate our method's strong ability to extract causal driving signals from diverse real-world datasets spanning neuroscience, genomics, fluid dynamics, and physiology.
comment: 8 pages, 5 figures
☆ Fourier Sensitivity and Regularization of Computer Vision Models
Recent work has empirically shown that deep neural networks latch on to the Fourier statistics of training data and show increased sensitivity to Fourier-basis directions in the input. Understanding and modifying this Fourier-sensitivity of computer vision models may help improve their robustness. Hence, in this paper we study the frequency sensitivity characteristics of deep neural networks using a principled approach. We first propose a basis trick, proving that unitary transformations of the input-gradient of a function can be used to compute its gradient in the basis induced by the transformation. Using this result, we propose a general measure of any differentiable model's Fourier-sensitivity using the unitary Fourier-transform of its input-gradient. When applied to deep neural networks, we find that computer vision models are consistently sensitive to particular frequencies dependent on the dataset, training method and architecture. Based on this measure, we further propose a Fourier-regularization framework to modify the Fourier-sensitivities and frequency bias of models. Using our proposed regularizer-family, we demonstrate that deep neural networks obtain improved classification accuracy on robustness evaluations.
comment: Published in TMLR, https://openreview.net/forum?id=VmTYgjYloM
☆ An Analysis of Classification Approaches for Hit Song Prediction using Engineered Metadata Features with Lyrics and Audio Features
Hit song prediction, one of the emerging fields in music information retrieval (MIR), remains a considerable challenge. Being able to understand what makes a given song a hit is clearly beneficial to the whole music industry. Previous approaches to hit song prediction have focused on using audio features of a record. This study aims to improve the prediction result of the top 10 hits among Billboard Hot 100 songs using more alternative metadata, including song audio features provided by Spotify, song lyrics, and novel metadata-based features (title topic, popularity continuity and genre class). Five machine learning approaches are applied, including: k-nearest neighbours, Naive Bayes, Random Forest, Logistic Regression and Multilayer Perceptron. Our results show that Random Forest (RF) and Logistic Regression (LR) with all features (including novel features, song audio features and lyrics features) outperforms other models, achieving 89.1% and 87.2% accuracy, and 0.91 and 0.93 AUC, respectively. Our findings also demonstrate the utility of our novel music metadata features, which contributed most to the models' discriminative performance.
☆ DNN Explanation for Safety Analysis: an Empirical Evaluation of Clustering-based Approaches
The adoption of deep neural networks (DNNs) in safety-critical contexts is often prevented by the lack of effective means to explain their results, especially when they are erroneous. In our previous work, we proposed a white-box approach (HUDD) and a black-box approach (SAFE) to automatically characterize DNN failures. They both identify clusters of similar images from a potentially large set of images leading to DNN failures. However, the analysis pipelines for HUDD and SAFE were instantiated in specific ways according to common practices, deferring the analysis of other pipelines to future work. In this paper, we report on an empirical evaluation of 99 different pipelines for root cause analysis of DNN failures. They combine transfer learning, autoencoders, heatmaps of neuron relevance, dimensionality reduction techniques, and different clustering algorithms. Our results show that the best pipeline combines transfer learning, DBSCAN, and UMAP. It leads to clusters almost exclusively capturing images of the same failure scenario, thus facilitating root cause analysis. Further, it generates distinct clusters for each root cause of failure, thus enabling engineers to detect all the unsafe scenarios. Interestingly, these results hold even for failure scenarios that are only observed in a small percentage of the failing images.
comment: 10 Tables, 14 Figures
☆ Auxiliary Learning as an Asymmetric Bargaining Game
Auxiliary learning is an effective method for enhancing the generalization capabilities of trained models, particularly when dealing with small datasets. However, this approach may present several difficulties: (i) optimizing multiple objectives can be more challenging, and (ii) how to balance the auxiliary tasks to best assist the main task is unclear. In this work, we propose a novel approach, named AuxiNash, for balancing tasks in auxiliary learning by formalizing the problem as generalized bargaining game with asymmetric task bargaining power. Furthermore, we describe an efficient procedure for learning the bargaining power of tasks based on their contribution to the performance of the main task and derive theoretical guarantees for its convergence. Finally, we evaluate AuxiNash on multiple multi-task benchmarks and find that it consistently outperforms competing methods.
☆ Company-as-Tribe: Company Financial Risk Assessment on Tribe-Style Graph with Hierarchical Graph Neural Networks KDD2022
Company financial risk is ubiquitous and early risk assessment for listed companies can avoid considerable losses. Traditional methods mainly focus on the financial statements of companies and lack the complex relationships among them. However, the financial statements are often biased and lagged, making it difficult to identify risks accurately and timely. To address the challenges, we redefine the problem as \textbf{company financial risk assessment on tribe-style graph} by taking each listed company and its shareholders as a tribe and leveraging financial news to build inter-tribe connections. Such tribe-style graphs present different patterns to distinguish risky companies from normal ones. However, most nodes in the tribe-style graph lack attributes, making it difficult to directly adopt existing graph learning methods (e.g., Graph Neural Networks(GNNs)). In this paper, we propose a novel Hierarchical Graph Neural Network (TH-GNN) for Tribe-style graphs via two levels, with the first level to encode the structure pattern of the tribes with contrastive learning, and the second level to diffuse information based on the inter-tribe relations, achieving effective and efficient risk assessment. Extensive experiments on the real-world company dataset show that our method achieves significant improvements on financial risk assessment over previous competing methods. Also, the extensive ablation studies and visualization comprehensively show the effectiveness of our method.
comment: accepted by SIGKDD2022
☆ Robust Linear Regression: Gradient-descent, Early-stopping, and Beyond
In this work we study the robustness to adversarial attacks, of early-stopping strategies on gradient-descent (GD) methods for linear regression. More precisely, we show that early-stopped GD is optimally robust (up to an absolute constant) against Euclidean-norm adversarial attacks. However, we show that this strategy can be arbitrarily sub-optimal in the case of general Mahalanobis attacks. This observation is compatible with recent findings in the case of classification~\cite{Vardi2022GradientMP} that show that GD provably converges to non-robust models. To alleviate this issue, we propose to apply instead a GD scheme on a transformation of the data adapted to the attack. This data transformation amounts to apply feature-depending learning rates and we show that this modified GD is able to handle any Mahalanobis attack, as well as more general attacks under some conditions. Unfortunately, choosing such adapted transformations can be hard for general attacks. To the rescue, we design a simple and tractable estimator whose adversarial risk is optimal up to within a multiplicative constant of 1.1124 in the population regime, and works for any norm.
☆ An investigation of challenges encountered when specifying training data and runtime monitors for safety critical ML applications
Context and motivation: The development and operation of critical software that contains machine learning (ML) models requires diligence and established processes. Especially the training data used during the development of ML models have major influences on the later behaviour of the system. Runtime monitors are used to provide guarantees for that behaviour. Question / problem: We see major uncertainty in how to specify training data and runtime monitoring for critical ML models and by this specifying the final functionality of the system. In this interview-based study we investigate the underlying challenges for these difficulties. Principal ideas/results: Based on ten interviews with practitioners who develop ML models for critical applications in the automotive and telecommunication sector, we identified 17 underlying challenges in 6 challenge groups that relate to the challenge of specifying training data and runtime monitoring. Contribution: The article provides a list of the identified underlying challenges related to the difficulties practitioners experience when specifying training data and runtime monitoring for ML models. Furthermore, interconnection between the challenges were found and based on these connections recommendation proposed to overcome the root causes for the challenges.
☆ GDOD: Effective Gradient Descent using Orthogonal Decomposition for Multi-Task Learning
Multi-task learning (MTL) aims at solving multiple related tasks simultaneously and has experienced rapid growth in recent years. However, MTL models often suffer from performance degeneration with negative transfer due to learning several tasks simultaneously. Some related work attributed the source of the problem is the conflicting gradients. In this case, it is needed to select useful gradient updates for all tasks carefully. To this end, we propose a novel optimization approach for MTL, named GDOD, which manipulates gradients of each task using an orthogonal basis decomposed from the span of all task gradients. GDOD decomposes gradients into task-shared and task-conflict components explicitly and adopts a general update rule for avoiding interference across all task gradients. This allows guiding the update directions depending on the task-shared components. Moreover, we prove the convergence of GDOD theoretically under both convex and non-convex assumptions. Experiment results on several multi-task datasets not only demonstrate the significant improvement of GDOD performed to existing MTL models but also prove that our algorithm outperforms state-of-the-art optimization methods in terms of AUC and Logloss metrics.
☆ Training with Mixed-Precision Floating-Point Assignments
When training deep neural networks, keeping all tensors in high precision (e.g., 32-bit or even 16-bit floats) is often wasteful. However, keeping all tensors in low precision (e.g., 8-bit floats) can lead to unacceptable accuracy loss. Hence, it is important to use a precision assignment -- a mapping from all tensors (arising in training) to precision levels (high or low) -- that keeps most of the tensors in low precision and leads to sufficiently accurate models. We provide a technique that explores this memory-accuracy tradeoff by generating precision assignments that (i) use less memory and (ii) lead to more accurate models at the same time, compared to the precision assignments considered by prior work in low-precision floating-point training. Our method typically provides > 2x memory reduction over a baseline precision assignment while preserving training accuracy, and gives further reductions by trading off accuracy. Compared to other baselines which sometimes cause training to diverge, our method provides similar or better memory reduction while avoiding divergence.
☆ Towards Learned Emulation of Interannual Water Isotopologue Variations in General Circulation Models
Simulating abundances of stable water isotopologues, i.e. molecules differing in their isotopic composition, within climate models allows for comparisons with proxy data and, thus, for testing hypotheses about past climate and validating climate models under varying climatic conditions. However, many models are run without explicitly simulating water isotopologues. We investigate the possibility to replace the explicit physics-based simulation of oxygen isotopic composition in precipitation using machine learning methods. These methods estimate isotopic composition at each time step for given fields of surface temperature and precipitation amount. We implement convolutional neural networks (CNNs) based on the successful UNet architecture and test whether a spherical network architecture outperforms the naive approach of treating Earth's latitude-longitude grid as a flat image. Conducting a case study on a last millennium run with the iHadCM3 climate model, we find that roughly 40\% of the temporal variance in the isotopic composition is explained by the emulations on interannual and monthly timescale, with spatially varying emulation quality. A modified version of the standard UNet architecture for flat images yields results that are equally good as the predictions by the spherical CNN. We test generalization to last millennium runs of other climate models and find that while the tested deep learning methods yield the best results on iHadCM3 data, the performance drops when predicting on other models and is comparable to simple pixel-wise linear regression. An extended choice of predictor variables and improving the robustness of learned climate--oxygen isotope relationships should be explored in future work.
☆ Learning Generalized Hybrid Proximity Representation for Image Recognition ICTAI 2022
Recently, deep metric learning techniques received attention, as the learned distance representations are useful to capture the similarity relationship among samples and further improve the performance of various of supervised or unsupervised learning tasks. We propose a novel supervised metric learning method that can learn the distance metrics in both geometric and probabilistic space for image recognition. In contrast to the previous metric learning methods which usually focus on learning the distance metrics in Euclidean space, our proposed method is able to learn better distance representation in a hybrid approach. To achieve this, we proposed a Generalized Hybrid Metric Loss (GHM-Loss) to learn the general hybrid proximity features from the image data by controlling the trade-off between geometric proximity and probabilistic proximity. To evaluate the effectiveness of our method, we first provide theoretical derivations and proofs of the proposed loss function, then we perform extensive experiments on two public datasets to show the advantage of our method compared to other state-of-the-art metric learning methods.
comment: The paper has been accepted by the IEEE ICTAI 2022
☆ A Data-Driven Modeling and Control Framework for Physics-Based Building Emulators
We present a data-driven modeling and control framework for physics-based building emulators. Our approach comprises: (a) Offline training of differentiable surrogate models that speed up model evaluations, provide cheap gradients, and have good predictive accuracy for the receding horizon in Model Predictive Control (MPC) and (b) Formulating and solving nonlinear building HVAC MPC problems. We extensively verify the modeling and control performance using multiple surrogate models and optimization frameworks for different available test cases in the Building Optimization Testing Framework (BOPTEST). The framework is compatible with other modeling techniques and customizable with different control formulations. The modularity makes the approach future-proof for test cases currently in development for physics-based building emulators and provides a path toward prototyping predictive controllers in large buildings.
☆ Sharp Variance-Dependent Bounds in Reinforcement Learning: Best of Both Worlds in Stochastic and Deterministic Environments
We study variance-dependent regret bounds for Markov decision processes (MDPs). Algorithms with variance-dependent regret guarantees can automatically exploit environments with low variance (e.g., enjoying constant regret on deterministic MDPs). The existing algorithms are either variance-independent or suboptimal. We first propose two new environment norms to characterize the fine-grained variance properties of the environment. For model-based methods, we design a variant of the MVP algorithm (Zhang et al., 2021a) and use new analysis techniques show to this algorithm enjoys variance-dependent bounds with respect to our proposed norms. In particular, this bound is simultaneously minimax optimal for both stochastic and deterministic MDPs, the first result of its kind. We further initiate the study on model-free algorithms with variance-dependent regret bounds by designing a reference-function-based algorithm with a novel capped-doubling reference update schedule. Lastly, we also provide lower bounds to complement our upper bounds.
comment: 43 pages, 1 figure
☆ A Survey of Explainable AI in Deep Visual Modeling: Methods and Metrics
Deep visual models have widespread applications in high-stake domains. Hence, their black-box nature is currently attracting a large interest of the research community. We present the first survey in Explainable AI that focuses on the methods and metrics for interpreting deep visual models. Covering the landmark contributions along the state-of-the-art, we not only provide a taxonomic organization of the existing techniques, but also excavate a range of evaluation metrics and collate them as measures of different properties of model explanations. Along the insightful discussion on the current trends, we also discuss the challenges and future avenues for this research direction.
comment: Short accessible survey (9pgs)
☆ Retiring $Δ$DP: New Distribution-Level Metrics for Demographic Parity
Demographic parity is the most widely recognized measure of group fairness in machine learning, which ensures equal treatment of different demographic groups. Numerous works aim to achieve demographic parity by pursuing the commonly used metric $\Delta DP$. Unfortunately, in this paper, we reveal that the fairness metric $\Delta DP$ can not precisely measure the violation of demographic parity, because it inherently has the following drawbacks: \textit{i)} zero-value $\Delta DP$ does not guarantee zero violation of demographic parity, \textit{ii)} $\Delta DP$ values can vary with different classification thresholds. To this end, we propose two new fairness metrics, \textsf{A}rea \textsf{B}etween \textsf{P}robability density function \textsf{C}urves (\textsf{ABPC}) and \textsf{A}rea \textsf{B}etween \textsf{C}umulative density function \textsf{C}urves (\textsf{ABCC}), to precisely measure the violation of demographic parity in distribution level. The new fairness metrics directly measure the difference between the distributions of the prediction probability for different demographic groups. Thus our proposed new metrics enjoy: \textit{i)} zero-value \textsf{ABCC}/\textsf{ABPC} guarantees zero violation of demographic parity; \textit{ii)} \textsf{ABCC}/\textsf{ABPC} guarantees demographic parity while the classification threshold adjusted. We further re-evaluate the existing fair models with our proposed fairness metrics and observe different fairness behaviors of those models under the new metrics.
comment: Under review
☆ Scaling laws for single-agent reinforcement learning
Recent work has shown that, in generative modeling, cross-entropy loss improves smoothly with model size and training compute, following a power law plus constant scaling law. One challenge in extending these results to reinforcement learning is that the main performance objective of interest, mean episode return, need not vary smoothly. To overcome this, we introduce *intrinsic performance*, a monotonic function of the return defined as the minimum compute required to achieve the given return across a family of models of different sizes. We find that, across a range of environments, intrinsic performance scales as a power law in model size and environment interactions. Consequently, as in generative modeling, the optimal model size scales as a power law in the training compute budget. Furthermore, we study how this relationship varies with the environment and with other properties of the training setup. In particular, using a toy MNIST-based environment, we show that varying the "horizon length" of the task mostly changes the coefficient but not the exponent of this relationship.
comment: 33 pages
☆ CMLCompiler: A Unified Compiler for Classical Machine Learning
Classical machine learning (CML) occupies nearly half of machine learning pipelines in production applications. Unfortunately, it fails to utilize the state-of-the-practice devices fully and performs poorly. Without a unified framework, the hybrid deployments of deep learning (DL) and CML also suffer from severe performance and portability issues. This paper presents the design of a unified compiler, called CMLCompiler, for CML inference. We propose two unified abstractions: operator representations and extended computational graphs. The CMLCompiler framework performs the conversion and graph optimization based on two unified abstractions, then outputs an optimized computational graph to DL compilers or frameworks. We implement CMLCompiler on TVM. The evaluation shows CMLCompiler's portability and superior performance. It achieves up to 4.38x speedup on CPU, 3.31x speedup on GPU, and 5.09x speedup on IoT devices, compared to the state-of-the-art solutions -- scikit-learn, intel sklearn, and hummingbird. Our performance of CML and DL mixed pipelines achieves up to 3.04x speedup compared with cross-framework implementations.
☆ Contrast and Clustering: Learning Neighborhood Pair Representation for Source-free Domain Adaptation
Domain adaptation has attracted a great deal of attention in the machine learning community, but it requires access to source data, which often raises concerns about data privacy. We are thus motivated to address these issues and propose a simple yet efficient method. This work treats domain adaptation as an unsupervised clustering problem and trains the target model without access to the source data. Specifically, we propose a loss function called contrast and clustering (CaC), where a positive pair term pulls neighbors belonging to the same class together in the feature space to form clusters, while a negative pair term pushes samples of different classes apart. In addition, extended neighbors are taken into account by querying the nearest neighbor indexes in the memory bank to mine for more valuable negative pairs. Extensive experiments on three common benchmarks, VisDA, Office-Home and Office-31, demonstrate that our method achieves state-of-the-art performance. The code will be made publicly available at https://github.com/yukilulu/CaC.
comment: conference paper
☆ Superhuman Fairness
The fairness of machine learning-based decisions has become an increasingly important focus in the design of supervised machine learning methods. Most fairness approaches optimize a specified trade-off between performance measure(s) (e.g., accuracy, log loss, or AUC) and fairness metric(s) (e.g., demographic parity, equalized odds). This begs the question: are the right performance-fairness trade-offs being specified? We instead re-cast fair machine learning as an imitation learning task by introducing superhuman fairness, which seeks to simultaneously outperform human decisions on multiple predictive performance and fairness measures. We demonstrate the benefits of this approach given suboptimal decisions.
☆ Revisiting Bellman Errors for Offline Model Selection
Offline model selection (OMS), that is, choosing the best policy from a set of many policies given only logged data, is crucial for applying offline RL in real-world settings. One idea that has been extensively explored is to select policies based on the mean squared Bellman error (MSBE) of the associated Q-functions. However, previous work has struggled to obtain adequate OMS performance with Bellman errors, leading many researchers to abandon the idea. Through theoretical and empirical analyses, we elucidate why previous work has seen pessimistic results with Bellman errors and identify conditions under which OMS algorithms based on Bellman errors will perform well. Moreover, we develop a new estimator of the MSBE that is more accurate than prior methods and obtains impressive OMS performance on diverse discrete control tasks, including Atari games. We open-source our data and code to enable researchers to conduct OMS experiments more easily.
comment: Main paper: 9 pages, 4 figures, 2 tables Main + Appendix: 31 pages, 11 figures, 2 tables For associated source code, please see https://github.com/jzitovsky/SBV
☆ Generating High Fidelity Synthetic Data via Coreset selection and Entropic Regularization NeurIPS 2022
Generative models have the ability to synthesize data points drawn from the data distribution, however, not all generated samples are high quality. In this paper, we propose using a combination of coresets selection methods and ``entropic regularization'' to select the highest fidelity samples. We leverage an Energy-Based Model which resembles a variational auto-encoder with an inference and generator model for which the latent prior is complexified by an energy-based model. In a semi-supervised learning scenario, we show that augmenting the labeled data-set, by adding our selected subset of samples, leads to better accuracy improvement rather than using all the synthetic samples.
comment: NeurIPS 2022 Workshop on Synthetic Data for Empowering ML Research
☆ Learning Topology-Preserving Data Representations ICLR 2023
We propose a method for learning topology-preserving data representations (dimensionality reduction). The method aims to provide topological similarity between the data manifold and its latent representation via enforcing the similarity in topological features (clusters, loops, 2D voids, etc.) and their localization. The core of the method is the minimization of the Representation Topology Divergence (RTD) between original high-dimensional data and low-dimensional representation in latent space. RTD minimization provides closeness in topological features with strong theoretical guarantees. We develop a scheme for RTD differentiation and apply it as a loss term for the autoencoder. The proposed method ``RTD-AE'' better preserves the global structure and topology of the data manifold than state-of-the-art competitors as measured by linear correlation, triplet distance ranking accuracy, and Wasserstein distance between persistence barcodes.
comment: 11th International Conference on Learning Representations (ICLR 2023)
♻ ☆ Limitations of Information-Theoretic Generalization Bounds for Gradient Descent Methods in Stochastic Convex Optimization ALT
To date, no "information-theoretic" frameworks for reasoning about generalization error have been shown to establish minimax rates for gradient descent in the setting of stochastic convex optimization. In this work, we consider the prospect of establishing such rates via several existing information-theoretic frameworks: input-output mutual information bounds, conditional mutual information bounds and variants, PAC-Bayes bounds, and recent conditional variants thereof. We prove that none of these bounds are able to establish minimax rates. We then consider a common tactic employed in studying gradient methods, whereby the final iterate is corrupted by Gaussian noise, producing a noisy "surrogate" algorithm. We prove that minimax rates cannot be established via the analysis of such surrogates. Our results suggest that new ideas are required to analyze gradient descent using information-theoretic techniques.
comment: 49 pages, 2 figures. To appear, Proc. International Conference on Algorithmic Learning Theory (ALT), 2023
♻ ☆ A Learning Based Hypothesis Test for Harmful Covariate Shift
The ability to quickly and accurately identify covariate shift at test time is a critical and often overlooked component of safe machine learning systems deployed in high-risk domains. While methods exist for detecting when predictions should not be made on out-of-distribution test examples, identifying distributional level differences between training and test time can help determine when a model should be removed from the deployment setting and retrained. In this work, we define harmful covariate shift (HCS) as a change in distribution that may weaken the generalization of a predictive model. To detect HCS, we use the discordance between an ensemble of classifiers trained to agree on training data and disagree on test data. We derive a loss function for training this ensemble and show that the disagreement rate and entropy represent powerful discriminative statistics for HCS. Empirically, we demonstrate the ability of our method to detect harmful covariate shift with statistical certainty on a variety of high-dimensional datasets. Across numerous domains and modalities, we show state-of-the-art performance compared to existing methods, particularly when the number of observed test samples is small.
♻ ☆ Stationary Kernels and Gaussian Processes on Lie Groups and their Homogeneous Spaces I: the compact case
Gaussian processes are arguably the most important model class in spatial statistics. They encode prior information about the modeled function and can be used for exact or approximate Bayesian inference. In many applications, particularly in physical sciences and engineering, but also in areas such as geostatistics and neuroscience, invariance to symmetries is one of the most fundamental forms of prior information one can consider. The invariance of a Gaussian process' covariance to such symmetries gives rise to the most natural generalization of the concept of stationarity to such spaces. In this work, we develop constructive and practical techniques for building stationary Gaussian processes on a very large class of non-Euclidean spaces arising in the context of symmetries. Our techniques make it possible to (i) calculate covariance kernels and (ii) sample from prior and posterior Gaussian processes defined on such spaces, both in a practical manner. This work is split into two parts, each involving different technical considerations: part I studies compact spaces, while part II studies non-compact spaces possessing certain structure. Our contributions make the non-Euclidean Gaussian process models we study compatible with well-understood computational techniques available in standard Gaussian process software packages, thereby making them accessible to practitioners.
♻ ☆ Causal Graph Discovery from Self and Mutually Exciting Time Series ICML 2021
We present a generalized linear structural causal model, coupled with a novel data-adaptive linear regularization, to recover causal directed acyclic graphs (DAGs) from time series. By leveraging a recently developed stochastic monotone Variational Inequality (VI) formulation, we cast the causal discovery problem as a general convex optimization. Furthermore, we develop a non-asymptotic recovery guarantee and quantifiable uncertainty by solving a linear program to establish confidence intervals for a wide range of non-linear monotone link functions. We validate our theoretical results and show the competitive performance of our method via extensive numerical experiments. Most importantly, we demonstrate the effectiveness of our approach in recovering highly interpretable causal DAGs over Sepsis Associated Derangements (SADs) while achieving comparable prediction performance to powerful ``black-box'' models such as XGBoost. Thus, the future adoption of our proposed method to conduct continuous surveillance of high-risk patients by clinicians is much more likely.
comment: See v2 for a previous workshop paper on Interpretable ML in Healthcare (IMLH) at ICML 2021, titled "Causal Graph Recovery for Sepsis-Associated Derangements via Interpretable Hawkes Networks". Also, see arXiv:2301.11336 for a short conference version with more experiments of our proposed method to learn "strict" DAGs
♻ ☆ Py-Feat: Python Facial Expression Analysis Toolbox
Studying facial expressions is a notoriously difficult endeavor. Recent advances in the field of affective computing have yielded impressive progress in automatically detecting facial expressions from pictures and videos. However, much of this work has yet to be widely disseminated in social science domains such as psychology. Current state of the art models require considerable domain expertise that is not traditionally incorporated into social science training programs. Furthermore, there is a notable absence of user-friendly and open-source software that provides a comprehensive set of tools and functions that support facial expression research. In this paper, we introduce Py-Feat, an open-source Python toolbox that provides support for detecting, preprocessing, analyzing, and visualizing facial expression data. Py-Feat makes it easy for domain experts to disseminate and benchmark computer vision models and also for end users to quickly process, analyze, and visualize face expression data. We hope this platform will facilitate increased use of facial expression data in human behavior research.
♻ ☆ Efficient Global Planning in Large MDPs via Stochastic Primal-Dual Optimization
We propose a new stochastic primal-dual optimization algorithm for planning in a large discounted Markov decision process with a generative model and linear function approximation. Assuming that the feature map approximately satisfies standard realizability and Bellman-closedness conditions and also that the feature vectors of all state-action pairs are representable as convex combinations of a small core set of state-action pairs, we show that our method outputs a near-optimal policy after a polynomial number of queries to the generative model. Our method is computationally efficient and comes with the major advantage that it outputs a single softmax policy that is compactly represented by a low-dimensional parameter vector, and does not need to execute computationally expensive local planning subroutines in runtime.
comment: 23 pages including reference and appendix
♻ ☆ Unsupervised Music Source Separation Using Differentiable Parametric Source Models
Supervised deep learning approaches to underdetermined audio source separation achieve state-of-the-art performance but require a dataset of mixtures along with their corresponding isolated source signals. Such datasets can be extremely costly to obtain for musical mixtures. This raises a need for unsupervised methods. We propose a novel unsupervised model-based deep learning approach to musical source separation. Each source is modelled with a differentiable parametric source-filter model. A neural network is trained to reconstruct the observed mixture as a sum of the sources by estimating the source models' parameters given their fundamental frequencies. At test time, soft masks are obtained from the synthesized source signals. The experimental evaluation on a vocal ensemble separation task shows that the proposed method outperforms learning-free methods based on nonnegative matrix factorization and a supervised deep learning baseline. Integrating domain knowledge in the form of source models into a data-driven method leads to high data efficiency: the proposed approach achieves good separation quality even when trained on less than three minutes of audio. This work makes powerful deep learning based separation usable in scenarios where training data with ground truth is expensive or nonexistent.
comment: Revised version of the submission
♻ ☆ Learning from many trajectories
We initiate a study of supervised learning from many independent sequences ("trajectories") of non-independent covariates, reflecting tasks in sequence modeling, control, and reinforcement learning. Conceptually, our multi-trajectory setup sits between two traditional settings in statistical learning theory: learning from independent examples and learning from a single auto-correlated sequence. Our conditions for efficient learning generalize the former setting--trajectories must be non-degenerate in ways that extend standard requirements for independent examples. Notably, we do not require that trajectories be ergodic, long, nor strictly stable. For linear least-squares regression, given $n$-dimensional examples produced by $m$ trajectories, each of length $T$, we observe a notable change in statistical efficiency as the number of trajectories increases from a few (namely $m \lesssim n$) to many (namely $m \gtrsim n$). Specifically, we establish that the worst-case error rate of this problem is $\Theta(n / m T)$ whenever $m \gtrsim n$. Meanwhile, when $m \lesssim n$, we establish a (sharp) lower bound of $\Omega(n^2 / m^2 T)$ on the worst-case error rate, realized by a simple, marginally unstable linear dynamical system. A key upshot is that, in domains where trajectories regularly reset, the error rate eventually behaves as if all of the examples were independent, drawn from their marginals. As a corollary of our analysis, we also improve guarantees for the linear system identification problem.
♻ ☆ A Unified Causal View of Domain Invariant Representation Learning
Machine learning methods can be unreliable when deployed in domains that differ from the domains on which they were trained. There are a wide range of proposals for mitigating this problem by learning representations that are ``invariant'' in some sense.However, these methods generally contradict each other, and none of them consistently improve performance on real-world domain shift benchmarks. There are two main questions that must be addressed to understand when, if ever, we should use each method. First, how does each ad hoc notion of ``invariance'' relate to the structure of real-world problems? And, second, when does learning invariant representations actually yield robust models? To address these issues, we introduce a broad formal notion of what it means for a real-world domain shift to admit invariant structure. Then, we characterize the causal structures that are compatible with this notion of invariance.With this in hand, we find conditions under which method-specific invariance notions correspond to real-world invariant structure, and we clarify the relationship between invariant structure and robustness to domain shifts. For both questions, we find that the true underlying causal structure of the data plays a critical role.
♻ ☆ Prioritizing Samples in Reinforcement Learning with Reducible Loss NeurIPS 2022
Most reinforcement learning algorithms take advantage of an experience replay buffer to repeatedly train on samples the agent has observed in the past. Not all samples carry the same amount of significance and simply assigning equal importance to each of the samples is a na\"ive strategy. In this paper, we propose a method to prioritize samples based on how much we can learn from a sample. We define the learn-ability of a sample as the steady decrease of the training loss associated with this sample over time. We develop an algorithm to prioritize samples with high learn-ability, while assigning lower priority to those that are hard-to-learn, typically caused by noise or stochasticity. We empirically show that our method is more robust than random sampling and also better than just prioritizing with respect to the training loss, i.e. the temporal difference loss, which is used in prioritized experience replay.
comment: DeepRL Workshop, NeurIPS 2022
♻ ☆ Video Influencers: Unboxing the Mystique
Influencer marketing has become a very popular tool to reach customers. Despite the rapid growth in influencer videos, there has been little research on the effectiveness of their constituent elements in explaining video engagement. We study YouTube influencers and analyze their unstructured video data across text, audio and images using a novel "interpretable deep learning" framework that accomplishes both goals of prediction and interpretation. Our prediction-based approach analyzes unstructured data and finds that "what is said" in words (text) is more influential than "how it is said" in imagery (images) followed by acoustics (audio). Our interpretation-based approach is implemented after completion of model prediction by analyzing the same source of unstructured data to measure importance attributed to the video elements. We eliminate several spurious and confounded relationships, and identify a smaller subset of theory-based relationships. We uncover novel findings that establish distinct effects for measures of shallow and deep engagement which are based on the dual-system framework of human thinking. Our approach is validated using simulated data, and we discuss the learnings from our findings for influencers and brands.
comment: 45 pages, Online Appendix
♻ ☆ Bayesian Calibration of Imperfect Computer Models using Physics-Informed Priors
We introduce a computational efficient data-driven framework suitable for quantifying the uncertainty in physical parameters and model formulation of computer models, represented by differential equations. We construct physics-informed priors, which are multi-output GP priors that encode the model's structure in the covariance function. This is extended into a fully Bayesian framework that quantifies the uncertainty of physical parameters and model predictions. Since physical models often are imperfect descriptions of the real process, we allow the model to deviate from the observed data by considering a discrepancy function. For inference, Hamiltonian Monte Carlo is used. Further, approximations for big data are developed that reduce the computational complexity from $\mathcal{O}(N^3)$ to $\mathcal{O}(N\cdot m^2),$ where $m \ll N.$ Our approach is demonstrated in simulation and real data case studies where the physics are described by time-dependent ODEs describe (cardiovascular models) and space-time dependent PDEs (heat equation). In the studies, it is shown that our modelling framework can recover the true parameters of the physical models in cases where 1) the reality is more complex than our modelling choice and 2) the data acquisition process is biased while also producing accurate predictions. Furthermore, it is demonstrated that our approach is computationally faster than traditional Bayesian calibration methods.
comment: 48 pages, 21 figures
♻ ☆ Forecasting COVID- 19 cases using Statistical Models and Ontology-based Semantic Modelling: A real time data analytics approach
SARS-COV-19 is the most prominent issue which many countries face today. The frequent changes in infections, recovered and deaths represents the dynamic nature of this pandemic. It is very crucial to predict the spreading rate of this virus for accurate decision making against fighting with the situation of getting infected through the virus, tracking and controlling the virus transmission in the community. We develop a prediction model using statistical time series models such as SARIMA and FBProphet to monitor the daily active, recovered and death cases of COVID-19 accurately. Then with the help of various details across each individual patient (like height, weight, gender etc.), we designed a set of rules using Semantic Web Rule Language and some mathematical models for dealing with COVID19 infected cases on an individual basis. After combining all the models, a COVID-19 Ontology is developed and performs various queries using SPARQL query on designed Ontology which accumulate the risk factors, provide appropriate diagnosis, precautions and preventive suggestions for COVID Patients. After comparing the performance of SARIMA and FBProphet, it is observed that the SARIMA model performs better in forecasting of COVID cases. On individual basis COVID case prediction, approx. 497 individual samples have been tested and classified into five different levels of COVID classes such as Having COVID, No COVID, High Risk COVID case, Medium to High Risk case, and Control needed case.
♻ ☆ STI: Turbocharge NLP Inference at the Edge via Elastic Pipelining ASPLOS'23
Natural Language Processing (NLP) inference is seeing increasing adoption by mobile applications, where on-device inference is desirable for crucially preserving user data privacy and avoiding network roundtrips. Yet, the unprecedented size of an NLP model stresses both latency and memory, creating a tension between the two key resources of a mobile device. To meet a target latency, holding the whole model in memory launches execution as soon as possible but increases one app's memory footprints by several times, limiting its benefits to only a few inferences before being recycled by mobile memory management. On the other hand, loading the model from storage on demand incurs IO as long as a few seconds, far exceeding the delay range satisfying to a user; pipelining layerwise model loading and execution does not hide IO either, due to the high skewness between IO and computation delays. To this end, we propose Speedy Transformer Inference (STI). Built on the key idea of maximizing IO/compute resource utilization on the most important parts of a model, STI reconciles the latency v.s. memory tension via two novel techniques. First, model sharding. STI manages model parameters as independently tunable shards, and profiles their importance to accuracy. Second, elastic pipeline planning with a preload buffer. STI instantiates an IO/compute pipeline and uses a small buffer for preload shards to bootstrap execution without stalling at early stages; it judiciously selects, tunes, and assembles shards per their importance for resource-elastic execution, maximizing inference accuracy. Atop two commodity SoCs, we build STI and evaluate it against a wide range of NLP tasks, under a practical range of target latencies, and on both CPU and GPU. We demonstrate that STI delivers high accuracies with 1-2 orders of magnitude lower memory, outperforming competitive baselines.
comment: ASPLOS'23
♻ ☆ Learning useful representations for shifting tasks and distributions
Does the dominant approach to learn representations (as a side effect of optimizing an expected cost for a single training distribution) remain a good approach when we are dealing with multiple distributions? Our thesis is that such scenarios are better served by representations that are richer than those obtained with a single optimization episode. We support this thesis with simple theoretical arguments and with experiments utilizing an apparently na\"{\i}ve ensembling technique: concatenating the representations obtained from multiple training episodes using the same data, model, algorithm, and hyper-parameters, but different random seeds. These independently trained networks perform similarly. Yet, in a number of scenarios involving new distributions, the concatenated representation performs substantially better than an equivalently sized network trained with a single training run. This proves that the representations constructed by multiple training episodes are in fact different. Although their concatenation carries little additional information about the training task under the training distribution, it becomes substantially more informative when tasks or distributions change. Meanwhile, a single training episode is unlikely to yield such a redundant representation because the optimization process has no reason to accumulate features that do not incrementally improve the training performance.
comment: 20 pages
♻ ☆ Diffuser: Efficient Transformers with Multi-hop Attention Diffusion for Long Sequences
Efficient Transformers have been developed for long sequence modeling, due to their subquadratic memory and time complexity. Sparse Transformer is a popular approach to improving the efficiency of Transformers by restricting self-attention to locations specified by the predefined sparse patterns. However, leveraging sparsity may sacrifice expressiveness compared to full-attention, when important token correlations are multiple hops away. To combine advantages of both the efficiency of sparse transformer and the expressiveness of full-attention Transformer, we propose \textit{Diffuser}, a new state-of-the-art efficient Transformer. Diffuser incorporates all token interactions within one attention layer while maintaining low computation and memory costs. The key idea is to expand the receptive field of sparse attention using Attention Diffusion, which computes multi-hop token correlations based on all paths between corresponding disconnected tokens, besides attention among neighboring tokens. Theoretically, we show the expressiveness of Diffuser as a universal sequence approximator for sequence-to-sequence modeling, and investigate its ability to approximate full-attention by analyzing the graph expander property from the spectral perspective. Experimentally, we investigate the effectiveness of Diffuser with extensive evaluations, including language modeling, image modeling, and Long Range Arena (LRA). Evaluation results show that Diffuser achieves improvements by an average of 0.94% on text classification tasks and 2.30% on LRA, with 1.67$\times$ memory savings compared to state-of-the-art benchmarks, which demonstrates superior performance of Diffuser in both expressiveness and efficiency aspects.
♻ ☆ Direct-Effect Risk Minimization for Domain Generalization
We study the problem of out-of-distribution (o.o.d.) generalization where spurious correlations of attributes vary across training and test domains. This is known as the problem of correlation shift and has posed concerns on the reliability of machine learning. In this work, we introduce the concepts of direct and indirect effects from causal inference to the domain generalization problem. We argue that models that learn direct effects minimize the worst-case risk across correlation-shifted domains. To eliminate the indirect effects, our algorithm consists of two stages: in the first stage, we learn an indirect-effect representation by minimizing the prediction error of domain labels using the representation and the class label; in the second stage, we remove the indirect effects learned in the first stage by matching each data with another data of similar indirect-effect representation but of different class label. We also propose a new model selection method by matching the validation set in the same way, which is shown to improve the generalization performance of existing models on correlation-shifted datasets. Experiments on 5 correlation-shifted datasets and the DomainBed benchmark verify the effectiveness of our approach.
♻ ☆ LieGG: Studying Learned Lie Group Generators
Symmetries built into a neural network have appeared to be very beneficial for a wide range of tasks as it saves the data to learn them. We depart from the position that when symmetries are not built into a model a priori, it is advantageous for robust networks to learn symmetries directly from the data to fit a task function. In this paper, we present a method to extract symmetries learned by a neural network and to evaluate the degree to which a network is invariant to them. With our method, we are able to explicitly retrieve learned invariances in a form of the generators of corresponding Lie-groups without prior knowledge of symmetries in the data. We use the proposed method to study how symmetrical properties depend on a neural network's parameterization and configuration. We found that the ability of a network to learn symmetries generalizes over a range of architectures. However, the quality of learned symmetries depends on the depth and the number of parameters.
♻ ☆ PINCH: An Adversarial Extraction Attack Framework for Deep Learning Models
Adversarial extraction attacks constitute an insidious threat against Deep Learning (DL) models in-which an adversary aims to steal the architecture, parameters, and hyper-parameters of a targeted DL model. Existing extraction attack literature have observed varying levels of attack success for different DL models and datasets, yet the underlying cause(s) behind their susceptibility often remain unclear, and would help facilitate creating secure DL systems. In this paper we present PINCH: an efficient and automated extraction attack framework capable of designing, deploying, and analyzing extraction attack scenarios across heterogeneous hardware platforms. Using PINCH, we perform extensive experimental evaluation of extraction attacks against 21 model architectures to explore new extraction attack scenarios and further attack staging. Our findings show (1) key extraction characteristics whereby particular model configurations exhibit strong resilience against specific attacks, (2) even partial extraction success enables further staging for other adversarial attacks, and (3) equivalent stolen models uncover differences in expressive power, yet exhibit similar captured knowledge.
comment: 19 pages, 13 figures, 5 tables
♻ ☆ Strategyproof Decision-Making in Panel Data Settings and Beyond
We propose a framework for decision-making in the presence of strategic agents with panel data, a standard setting in econometrics and statistics where one gets noisy, repeated measurements of multiple units. We consider a setup where there is a pre-intervention period, when the principal observes the outcomes of each unit, after which the principal uses these observations to assign a treatment to each unit. Our model can be thought of as a generalization of the synthetic controls and synthetic interventions frameworks, where units (or agents) may strategically manipulate pre-intervention outcomes to receive a more desirable intervention. We identify necessary and sufficient conditions under which a strategyproof mechanism that assigns interventions in the post-intervention period exists. Under a latent factor model assumption, we show that whenever a strategyproof mechanism exists, there is one with a simple closed form. In the setting where there is a single treatment and control (i.e., no other interventions), we establish that there is always a strategyproof mechanism, and provide an algorithm for learning such a mechanism. For the setting of multiple interventions, we provide an algorithm for learning a strategyproof mechanism, if there exists a sufficiently large gap in rewards between the different interventions. Finally, we empirically evaluate our model using real-world panel data collected from product sales over 18 months. We find that our methods compare favorably to baselines which do not take strategic interactions into consideration -- even in the presence of model misspecification. Along the way, we prove impossibility results for multi-class strategic classification, which may be of independent interest.
♻ ☆ Sequential Kernelized Independence Testing
Independence testing is a fundamental and classical statistical problem that has been extensively studied in the batch setting when one fixes the sample size before collecting data. However, practitioners often prefer procedures that adapt to the complexity of a problem at hand instead of setting sample size in advance. Ideally, such procedures should (a) allow stopping earlier on easy tasks (and later on harder tasks), hence making better use of available resources, and (b) continuously monitor the data and efficiently incorporate statistical evidence after collecting new data, while controlling the false alarm rate. It is well known that classical batch tests are not tailored for streaming data settings: valid inference after data peeking requires correcting for multiple testing but such corrections generally result in low power. Following the principle of testing by betting, we design sequential kernelized independence tests (SKITs) that overcome such shortcomings. We exemplify our broad framework using bets inspired by kernelized dependence measures, e.g, the Hilbert-Schmidt independence criterion. Our test is valid under non-i.i.d. time-varying settings, for which there exist no batch tests. We demonstrate the power of our approaches on both simulated and real data.
♻ ☆ What can be learnt with wide convolutional neural networks?
Understanding how convolutional neural networks (CNNs) can efficiently learn high-dimensional functions remains a fundamental challenge. A popular belief is that these models harness the local and hierarchical structure of natural data such as images. Yet, we lack a quantitative understanding of how such structure affects performance, e.g. the rate of decay of the generalisation error with the number of training samples. In this paper, we study deep CNNs in the kernel regime. First, we show that the spectrum of the corresponding kernel inherits the hierarchical structure of the network, and we characterise its asymptotics. Then, we use this result together with generalisation bounds to prove that deep CNNs adapt to the spatial scale of the target function. In particular, we find that if the target function depends on low-dimensional subsets of adjacent input variables, then the rate of decay of the error is controlled by the effective dimensionality of these subsets. Conversely, if the target function depends on the full set of input variables, then the error rate is inversely proportional to the input dimension. We conclude by computing the rate when a deep CNN is trained on the output of another deep CNN with randomly-initialised parameters. Interestingly, we find that, despite their hierarchical structure, the functions generated by deep CNNs are too rich to be efficiently learnable in high dimension.
♻ ☆ Neural Continuous-Discrete State Space Models for Irregularly-Sampled Time Series
Learning accurate predictive models of real-world dynamic phenomena (e.g., climate, biological) remains a challenging task. One key issue is that the data generated by both natural and artificial processes often comprise time series that are irregularly sampled and/or contain missing observations. In this work, we propose the Neural Continuous-Discrete State Space Model (NCDSSM) for continuous-time modeling of time series through discrete-time observations. NCDSSM employs auxiliary variables to disentangle recognition from dynamics, thus requiring amortized inference only for the auxiliary variables. Leveraging techniques from continuous-discrete filtering theory, we demonstrate how to perform accurate Bayesian inference for the dynamic states. We propose three flexible parameterizations of the latent dynamics and an efficient training objective that marginalizes the dynamic states during inference. Empirical results on multiple benchmark datasets across various domains show improved imputation and forecasting performance of NCDSSM over existing models.
♻ ☆ What I Cannot Predict, I Do Not Understand: A Human-Centered Evaluation Framework for Explainability Methods
A multitude of explainability methods and associated fidelity performance metrics have been proposed to help better understand how modern AI systems make decisions. However, much of the current work has remained theoretical -- without much consideration for the human end-user. In particular, it is not yet known (1) how useful current explainability methods are in practice for more real-world scenarios and (2) how well associated performance metrics accurately predict how much knowledge individual explanations contribute to a human end-user trying to understand the inner-workings of the system. To fill this gap, we conducted psychophysics experiments at scale to evaluate the ability of human participants to leverage representative attribution methods for understanding the behavior of different image classifiers representing three real-world scenarios: identifying bias in an AI system, characterizing the visual strategy it uses for tasks that are too difficult for an untrained non-expert human observer as well as understanding its failure cases. Our results demonstrate that the degree to which individual attribution methods help human participants better understand an AI system varied widely across these scenarios. This suggests a critical need for the field to move past quantitative improvements of current attribution methods towards the development of complementary approaches that provide qualitatively different sources of information to human end-users.
♻ ☆ Normalizing Flows for Interventional Density Estimation
Existing machine learning methods for causal inference usually estimate quantities expressed via the mean of potential outcomes (e.g., average treatment effect). However, such quantities do not capture the full information about the distribution of potential outcomes. In this work, we estimate the density of potential outcomes after interventions from observational data. For this, we propose a novel, fully-parametric deep learning method called Interventional Normalizing Flows. Specifically, we combine two normalizing flows, namely (i) a teacher flow for estimating nuisance parameters and (ii) a student flow for a parametric estimation of the density of potential outcomes. We further develop a tractable optimization objective based on a one-step bias correction for an efficient and doubly robust estimation of the student flow parameters. As a result our Interventional Normalizing Flows offer a properly normalized density estimator. Across various experiments, we demonstrate that our Interventional Normalizing Flows are expressive and highly effective, and scale well with both sample size and high-dimensional confounding. To the best of our knowledge, our Interventional Normalizing Flows are the first fully-parametric, deep learning method for density estimation of potential outcomes.
♻ ☆ Antenna Array Calibration Via Gaussian Process Models
Antenna array calibration is necessary to maintain the high fidelity of beam patterns across a wide range of advanced antenna systems and to ensure channel reciprocity in time division duplexing schemes. Despite the continuous development in this area, most existing solutions are optimised for specific radio architectures, require standardised over-the-air data transmission, or serve as extensions of conventional methods. The diversity of communication protocols and hardware creates a problematic case, since this diversity requires to design or update the calibration procedures for each new advanced antenna system. In this study, we formulate antenna calibration in an alternative way, namely as a task of functional approximation, and address it via Bayesian machine learning. Our contributions are three-fold. Firstly, we define a parameter space, based on near-field measurements, that captures the underlying hardware impairments corresponding to each radiating element, their positional offsets, as well as the mutual coupling effects between antenna elements. Secondly, Gaussian process regression is used to form models from a sparse set of the aforementioned near-field data. Once deployed, the learned non-parametric models effectively serve to continuously transform the beamforming weights of the system, resulting in corrected beam patterns. Lastly, we demonstrate the viability of the described methodology for both digital and analog beamforming antenna arrays of different scales and discuss its further extension to support real-time operation with dynamic hardware impairments.
comment: International ITG 26th Workshop on Smart Antennas and 13th Conference on Systems, Communications, and Coding
♻ ☆ Deep Reinforcement Learning for Cryptocurrency Trading: Practical Approach to Address Backtest Overfitting
Designing profitable and reliable trading strategies is challenging in the highly volatile cryptocurrency market. Existing works applied deep reinforcement learning methods and optimistically reported increased profits in backtesting, which may suffer from the false positive issue due to overfitting. In this paper, we propose a practical approach to address backtest overfitting for cryptocurrency trading using deep reinforcement learning. First, we formulate the detection of backtest overfitting as a hypothesis test. Then, we train the DRL agents, estimate the probability of overfitting, and reject the overfitted agents, increasing the chance of good trading performance. Finally, on 10 cryptocurrencies over a testing period from 05/01/2022 to 06/27/2022 (during which the crypto market crashed two times), we show that the less overfitted deep reinforcement learning agents have a higher return than that of more overfitted agents, an equal weight strategy, and the S&P DBM Index (market benchmark), offering confidence in possible deployment to a real market.
♻ ☆ Revisiting Hyperparameter Tuning with Differential Privacy NeurIPS'22
Hyperparameter tuning is a common practice in the application of machine learning but is a typically ignored aspect in the literature on privacy-preserving machine learning due to its negative effect on the overall privacy parameter. In this paper, we aim to tackle this fundamental yet challenging problem by providing an effective hyperparameter tuning framework with differential privacy. The proposed method allows us to adopt a broader hyperparameter search space and even to perform a grid search over the whole space, since its privacy loss parameter is independent of the number of hyperparameter candidates. Interestingly, it instead correlates with the utility gained from hyperparameter searching, revealing an explicit and mandatory trade-off between privacy and utility. Theoretically, we show that its additional privacy loss bound incurred by hyperparameter tuning is upper-bounded by the squared root of the gained utility. However, we note that the additional privacy loss bound would empirically scale like a squared root of the logarithm of the utility term, benefiting from the design of doubling step.
comment: ML Safety Workshop of NeurIPS'22 Accepted Paper
♻ ☆ An SDE for Modeling SAM: Theory and Insights
We study the SAM (Sharpness-Aware Minimization) optimizer which has recently attracted a lot of interest due to its increased performance over more classical variants of stochastic gradient descent. Our main contribution is the derivation of continuous-time models (in the form of SDEs) for SAM and two of its variants, both for the full-batch and mini-batch settings. We demonstrate that these SDEs are rigorous approximations of the real discrete-time algorithms (in a weak sense, scaling linearly with the step size). Using these models, we then offer an explanation of why SAM prefers flat minima over sharp ones~--~by showing that it minimizes an implicitly regularized loss with a Hessian-dependent noise structure. Finally, we prove that perhaps unexpectedly SAM is attracted to saddle points under some realistic conditions. Our theoretical results are supported by detailed experiments.
♻ ☆ An $l_1$-oracle inequality for the Lasso in high-dimensional mixtures of experts models
Mixtures of experts (MoE) models are a popular framework for modeling heterogeneity in data, for both regression and classification problems in statistics and machine learning, due to their flexibility and the abundance of available statistical estimation and model choice tools. Such flexibility comes from allowing the mixture weights (or gating functions) in the MoE model to depend on the explanatory variables, along with the experts (or component densities). This permits the modeling of data arising from more complex data generating processes when compared to the classical finite mixtures and finite mixtures of regression models, whose mixing parameters are independent of the covariates. The use of MoE models in a high-dimensional setting, when the number of explanatory variables can be much larger than the sample size, is challenging from a computational point of view, and in particular from a theoretical point of view, where the literature is still lacking results for dealing with the curse of dimensionality, for both the statistical estimation and feature selection problems. We consider the finite MoE model with soft-max gating functions and Gaussian experts for high-dimensional regression on heterogeneous data, and its $l_1$-regularized estimation via the Lasso. We focus on the Lasso estimation properties rather than its feature selection properties. We provide a lower bound on the regularization parameter of the Lasso function that ensures an $l_1$-oracle inequality satisfied by the Lasso estimator according to the Kullback--Leibler loss.
comment: Added more explanations
♻ ☆ Can Persistent Homology provide an efficient alternative for Evaluation of Knowledge Graph Completion Methods? WWW'23
In this paper we present a novel method, $\textit{Knowledge Persistence}$ ($\mathcal{KP}$), for faster evaluation of Knowledge Graph (KG) completion approaches. Current ranking-based evaluation is quadratic in the size of the KG, leading to long evaluation times and consequently a high carbon footprint. $\mathcal{KP}$ addresses this by representing the topology of the KG completion methods through the lens of topological data analysis, concretely using persistent homology. The characteristics of persistent homology allow $\mathcal{KP}$ to evaluate the quality of the KG completion looking only at a fraction of the data. Experimental results on standard datasets show that the proposed metric is highly correlated with ranking metrics (Hits@N, MR, MRR). Performance evaluation shows that $\mathcal{KP}$ is computationally efficient: In some cases, the evaluation time (validation+test) of a KG completion method has been reduced from 18 hours (using Hits@10) to 27 seconds (using $\mathcal{KP}$), and on average (across methods & data) reduces the evaluation time (validation+test) by $\approx$ $\textbf{99.96}\%$.
comment: To appear in proceedings of The Web Conference 2023 (WWW'23)
♻ ☆ OPT-GAN: A Broad-Spectrum Global Optimizer for Black-box Problems by Learning Distribution
Black-box optimization (BBO) algorithms are concerned with finding the best solutions for problems with missing analytical details. Most classical methods for such problems are based on strong and fixed a priori assumptions, such as Gaussianity. However, the complex real-world problems, especially when the global optimum is desired, could be very far from the a priori assumptions because of their diversities, causing unexpected obstacles. In this study, we propose a generative adversarial net-based broad-spectrum global optimizer (OPT-GAN) which estimates the distribution of optimum gradually, with strategies to balance exploration-exploitation trade-off. It has potential to better adapt to the regularity and structure of diversified landscapes than other methods with fixed prior, e.g., Gaussian assumption or separability. Experiments on diverse BBO benchmarks and high dimensional real world applications exhibit that OPT-GAN outperforms other traditional and neural net-based BBO algorithms.
♻ ☆ Neural parameter calibration for large-scale multi-agent models
Computational models have become a powerful tool in the quantitative sciences to understand the behaviour of complex systems that evolve in time. However, they often contain a potentially large number of free parameters whose values cannot be obtained from theory but need to be inferred from data. This is especially the case for models in the social sciences, economics, or computational epidemiology. Yet many current parameter estimation methods are mathematically involved and computationally slow to run. In this paper we present a computationally simple and fast method to retrieve accurate probability densities for model parameters using neural differential equations. We present a pipeline comprising multi-agent models acting as forward solvers for systems of ordinary or stochastic differential equations, and a neural network to then extract parameters from the data generated by the model. The two combined create a powerful tool that can quickly estimate densities on model parameters, even for very large systems. We demonstrate the method on synthetic time series data of the SIR model of the spread of infection, and perform an in-depth analysis of the Harris-Wilson model of economic activity on a network, representing a non-convex problem. For the latter, we apply our method both to synthetic data and to data of economic activity across Greater London. We find that our method calibrates the model orders of magnitude more accurately than a previous study of the same dataset using classical techniques, while running between 195 and 390 times faster.
♻ ☆ Automated speech- and text-based classification of neuropsychiatric conditions in a multidiagnostic setting
Speech patterns have been identified as potential diagnostic markers for neuropsychiatric conditions. However, most studies only compare a single clinical group to healthy controls, whereas clinical practice often requires differentiating between multiple potential diagnoses (multiclass settings). To address this, we assembled a dataset of repeated recordings from 420 participants (67 with major depressive disorder, 106 with schizophrenia and 46 with autism, as well as matched controls), and tested the performance of a range of conventional machine learning models and advanced Transformer models on both binary and multiclass classification, based on voice and text features. While binary models performed comparably to previous research (F1 scores between 0.54-0.75 for autism spectrum disorder, ASD; 0.67-0.92 for major depressive disorder, MDD; and 0.71-0.83 for schizophrenia); when differentiating between multiple diagnostic groups performance decreased markedly (F1 scores between 0.35-0.44 for ASD, 0.57-0.75 for MDD, 0.15-0.66 for schizophrenia, and 0.38-0.52 macro F1). Combining voice and text-based models yielded increased performance, suggesting that they capture complementary diagnostic information. Our results indicate that models trained on binary classification may learn to rely on markers of generic differences between clinical and non-clinical populations, or markers of clinical features that overlap across conditions, rather than identifying markers specific to individual conditions. We provide recommendations for future research in the field, suggesting increased focus on developing larger transdiagnostic datasets that include more fine-grained clinical features, and that can support the development of models that better capture the complexity of neuropsychiatric conditions and naturalistic diagnostic assessment.
comment: 24 pages, 5 figures
♻ ☆ Meta-Learning via Classifier(-free) Diffusion Guidance
We introduce meta-learning algorithms that perform zero-shot weight-space adaptation of neural network models to unseen tasks. Our methods repurpose the popular generative image synthesis techniques of natural language guidance and diffusion models to generate neural network weights adapted for tasks. We first train an unconditional generative hypernetwork model to produce neural network weights; then we train a second "guidance" model that, given a natural language task description, traverses the hypernetwork latent space to find high-performance task-adapted weights in a zero-shot manner. We explore two alternative approaches for latent space guidance: "HyperCLIP"-based classifier guidance and a conditional Hypernetwork Latent Diffusion Model ("HyperLDM"), which we show to benefit from the classifier-free guidance technique common in image generation. Finally, we demonstrate that our approaches outperform existing multi-task and meta-learning methods in a series of zero-shot learning experiments on our Meta-VQA dataset.
♻ ☆ Accelerating Material Design with the Generative Toolkit for Scientific Discovery
With the growing availability of data within various scientific domains, generative models hold enormous potential to accelerate scientific discovery. They harness powerful representations learned from datasets to speed up the formulation of novel hypotheses with the potential to impact material discovery broadly. We present the Generative Toolkit for Scientific Discovery (GT4SD). This extensible open-source library enables scientists, developers, and researchers to train and use state-of-the-art generative models to accelerate scientific discovery focused on material design.
comment: 15 pages, 2 figures
♻ ☆ Optimal precision for GANs
Many deep generative models are defined as a push-forward of a Gaussian measure by a continuous generator, such as Generative Adversarial Networks (GANs) or Variational Auto-Encoders (VAEs). This work explores the latent space of such deep generative models. A key issue with these models is their tendency to output samples outside of the support of the target distribution when learning disconnected distributions. We investigate the relationship between the performance of these models and the geometry of their latent space. Building on recent developments in geometric measure theory, we prove a sufficient condition for optimality in the case where the dimension of the latent space is larger than the number of modes. Through experiments on GANs, we demonstrate the validity of our theoretical results and gain new insights into the latent space geometry of these models. Additionally, we propose a truncation method that enforces a simplicial cluster structure in the latent space and improves the performance of GANs.
♻ ☆ DoubleML -- An Object-Oriented Implementation of Double Machine Learning in R
The R package DoubleML implements the double/debiased machine learning framework of Chernozhukov et al. (2018). It provides functionalities to estimate parameters in causal models based on machine learning methods. The double machine learning framework consist of three key ingredients: Neyman orthogonality, high-quality machine learning estimation and sample splitting. Estimation of nuisance components can be performed by various state-of-the-art machine learning methods that are available in the mlr3 ecosystem. DoubleML makes it possible to perform inference in a variety of causal models, including partially linear and interactive regression models and their extensions to instrumental variable estimation. The object-oriented implementation of DoubleML enables a high flexibility for the model specification and makes it easily extendable. This paper serves as an introduction to the double machine learning framework and the R package DoubleML. In reproducible code examples with simulated and real data sets, we demonstrate how DoubleML users can perform valid inference based on machine learning methods.
comment: 42 pages, 8 Figures, 1 Table; Updated version for DoubleML > 0.5.0
♻ ☆ Cutting Plane Selection with Analytic Centers and Multiregression
Cutting planes are a crucial component of state-of-the-art mixed-integer programming solvers, with the choice of which subset of cuts to add being vital for solver performance. We propose new distance-based measures to qualify the value of a cut by quantifying the extent to which it separates relevant parts of the relaxed feasible set. For this purpose, we use the analytic centers of the relaxation polytope or of its optimal face, as well as alternative optimal solutions of the linear programming relaxation. We assess the impact of the choice of distance measure on root node performance and throughout the whole branch-and-bound tree, comparing our measures against those prevalent in the literature. Finally, by a multi-output regression, we predict the relative performance of each measure, using static features readily available before the separation process. Our results indicate that analytic center-based methods help to significantly reduce the number of branch-and-bound nodes needed to explore the search space and that our multiregression approach can further improve on any individual method.
♻ ☆ Fine-tuning or top-tuning? Transfer learning with pretrained features and fast kernel methods
The impressive performances of deep learning architectures is associated to massive increase of models complexity. Millions of parameters need be tuned, with training and inference time scaling accordingly. But is massive fine-tuning necessary? In this paper, focusing on image classification, we consider a simple transfer learning approach exploiting pretrained convolutional features as input for a fast kernel method. We refer to this approach as top-tuning, since only the kernel classifier is trained. By performing more than 2500 training processes we show that this top-tuning approach provides comparable accuracy w.r.t. fine-tuning, with a training time that is between one and two orders of magnitude smaller. These results suggest that top-tuning provides a useful alternative to fine-tuning in small/medium datasets, especially when training efficiency is crucial.
♻ ☆ Bayesian Learning for Neural Networks: an algorithmic survey
The last decade witnessed a growing interest in Bayesian learning. Yet, the technicality of the topic and the multitude of ingredients involved therein, besides the complexity of turning theory into practical implementations, limit the use of the Bayesian learning paradigm, preventing its widespread adoption across different fields and applications. This self-contained survey engages and introduces readers to the principles and algorithms of Bayesian Learning for Neural Networks. It provides an introduction to the topic from an accessible, practical-algorithmic perspective. Upon providing a general introduction to Bayesian Neural Networks, we discuss and present both standard and recent approaches for Bayesian inference, with an emphasis on solutions relying on Variational Inference and the use of Natural gradients. We also discuss the use of manifold optimization as a state-of-the-art approach to Bayesian learning. We examine the characteristic properties of all the discussed methods, and provide pseudo-codes for their implementation, paying attention to practical aspects, such as the computation of the gradients.
♻ ☆ SecGNN: Privacy-Preserving Graph Neural Network Training and Inference as a Cloud Service SC
Graphs are widely used to model the complex relationships among entities. As a powerful tool for graph analytics, graph neural networks (GNNs) have recently gained wide attention due to its end-to-end processing capabilities. With the proliferation of cloud computing, it is increasingly popular to deploy the services of complex and resource-intensive model training and inference in the cloud due to its prominent benefits. However, GNN training and inference services, if deployed in the cloud, will raise critical privacy concerns about the information-rich and proprietary graph data (and the resulting model). While there has been some work on secure neural network training and inference, they all focus on convolutional neural networks handling images and text rather than complex graph data with rich structural information. In this paper, we design, implement, and evaluate SecGNN, the first system supporting privacy-preserving GNN training and inference services in the cloud. SecGNN is built from a synergy of insights on lightweight cryptography and machine learning techniques. We deeply examine the procedure of GNN training and inference, and devise a series of corresponding secure customized protocols to support the holistic computation. Extensive experiments demonstrate that SecGNN achieves comparable plaintext training and inference accuracy, with promising performance.
comment: Accepted in IEEE Transactions on Services Computing (TSC)
♻ ☆ Double Sampling Randomized Smoothing ICML 2022
Neural networks (NNs) are known to be vulnerable against adversarial perturbations, and thus there is a line of work aiming to provide robustness certification for NNs, such as randomized smoothing, which samples smoothing noises from a certain distribution to certify the robustness for a smoothed classifier. However, as shown by previous work, the certified robust radius in randomized smoothing suffers from scaling to large datasets ("curse of dimensionality"). To overcome this hurdle, we propose a Double Sampling Randomized Smoothing (DSRS) framework, which exploits the sampled probability from an additional smoothing distribution to tighten the robustness certification of the previous smoothed classifier. Theoretically, under mild assumptions, we prove that DSRS can certify $\Theta(\sqrt d)$ robust radius under $\ell_2$ norm where $d$ is the input dimension, implying that DSRS may be able to break the curse of dimensionality of randomized smoothing. We instantiate DSRS for a generalized family of Gaussian smoothing and propose an efficient and sound computing method based on customized dual optimization considering sampling error. Extensive experiments on MNIST, CIFAR-10, and ImageNet verify our theory and show that DSRS certifies larger robust radii than existing baselines consistently under different settings. Code is available at https://github.com/llylly/DSRS.
comment: ICML 2022; minor typos fixed; minor data corrected on Page 42 (no influence on conclusions)
♻ ☆ Disentangling Model Multiplicity in Deep Learning
Model multiplicity is a well-known but poorly understood phenomenon that undermines the generalisation guarantees of machine learning models. It appears when two models with similar training-time performance differ in their predictions and real-world performance characteristics. This observed 'predictive' multiplicity (PM) also implies elusive differences in the internals of the models, their 'representational' multiplicity (RM). We introduce a conceptual and experimental setup for analysing RM by measuring activation similarity via singular vector canonical correlation analysis (SVCCA). We show that certain differences in training methods systematically result in larger RM than others and evaluate RM and PM over a finite sample as predictors for generalizability. We further correlate RM with PM measured by the variance in i.i.d. and out-of-distribution test predictions in four standard image data sets. Finally, instead of attempting to eliminate RM, we call for its systematic measurement and maximal exposure.
comment: 13 pages, 6 figures
♻ ☆ Self-Compressing Neural Networks
This work focuses on reducing neural network size, which is a major driver of neural network execution time, power consumption, bandwidth, and memory footprint. A key challenge is to reduce size in a manner that can be exploited readily for efficient training and inference without the need for specialized hardware. We propose Self-Compression: a simple, general method that simultaneously achieves two goals: (1) removing redundant weights, and (2) reducing the number of bits required to represent the remaining weights. This is achieved using a generalized loss function to minimize overall network size. In our experiments we demonstrate floating point accuracy with as few as 3% of the bits and 18% of the weights remaining in the network.
comment: Accepted submission to 2023 DL-Hardware Co-Design for AI Acceleration
♻ ☆ FreeMatch: Self-adaptive Thresholding for Semi-supervised Learning ICLR 2023
Semi-supervised Learning (SSL) has witnessed great success owing to the impressive performances brought by various methods based on pseudo labeling and consistency regularization. However, we argue that existing methods might fail to utilize the unlabeled data more effectively since they either use a pre-defined / fixed threshold or an ad-hoc threshold adjusting scheme, resulting in inferior performance and slow convergence. We first analyze a motivating example to obtain intuitions on the relationship between the desirable threshold and model's learning status. Based on the analysis, we hence propose FreeMatch to adjust the confidence threshold in a self-adaptive manner according to the model's learning status. We further introduce a self-adaptive class fairness regularization penalty to encourage the model for diverse predictions during the early training stage. Extensive experiments indicate the superiority of FreeMatch especially when the labeled data are extremely rare. FreeMatch achieves 5.78%, 13.59%, and 1.28% error rate reduction over the latest state-of-the-art method FlexMatch on CIFAR-10 with 1 label per class, STL-10 with 4 labels per class, and ImageNet with 100 labels per class, respectively. Moreover, FreeMatch can also boost the performance of imbalanced SSL. The codes can be found at https://github.com/microsoft/Semi-supervised-learning.
comment: Accepted by ICLR 2023. Code: https://github.com/microsoft/Semi-supervised-learning
♻ ☆ Incorporating Knowledge into Document Summarization: an Application of Prefix-Tuning on GPT-2
Despite the great development of document summarization techniques nowadays, factual inconsistencies between the generated summaries and the original text still occur from time to time. This paper proposes a prefix-tuning-based approach that uses a set of trainable continuous prefix prompt together with discrete prompts to aid model generation, which makes a significant impact on both CNN/Daily Mail and XSum summaries generated using GPT-2. The improvements on fact preservation in the generated summaries indicates the effectiveness of adopting this prefix-tuning-based method in knowledge-enhanced document summarization, and also shows a great potential on other natural language processing tasks.
♻ ☆ A Scalable, Interpretable, Verifiable & Differentiable Logic Gate Convolutional Neural Network Architecture From Truth Tables
We propose $\mathcal{T}$ruth $\mathcal{T}$able net ($\mathcal{TT}$net), a novel Convolutional Neural Network (CNN) architecture that addresses, by design, the open challenges of interpretability, formal verification, and logic gate conversion. $\mathcal{TT}$net is built using CNNs' filters that are equivalent to tractable truth tables and that we call Learning Truth Table (LTT) blocks. The dual form of LTT blocks allows the truth tables to be easily trained with gradient descent and makes these CNNs easy to interpret, verify and infer. Specifically, $\mathcal{TT}$net is a deep CNN model that can be automatically represented, after post-training transformation, as a sum of Boolean decision trees, or as a sum of Disjunctive/Conjunctive Normal Form (DNF/CNF) formulas, or as a compact Boolean logic circuit. We demonstrate the effectiveness and scalability of $\mathcal{TT}$net on multiple datasets, showing comparable interpretability to decision trees, fast complete/sound formal verification, and scalable logic gate representation, all compared to state-of-the-art methods. We believe this work represents a step towards making CNNs more transparent and trustworthy for real-world critical applications.
♻ ☆ Discovery of Single Independent Latent Variable
Latent variable discovery is a central problem in data analysis with a broad range of applications in applied science. In this work, we consider data given as an invertible mixture of two statistically independent components, and assume that one of the components is observed while the other is hidden. Our goal is to recover the hidden component. For this purpose, we propose an autoencoder equipped with a discriminator. Unlike the standard nonlinear ICA problem, which was shown to be non-identifiable, in the special case of ICA we consider here, we show that our approach can recover the component of interest up to entropy-preserving transformation. We demonstrate the performance of the proposed approach on several datasets, including image synthesis, voice cloning, and fetal ECG extraction.
comment: Published as a conference paper at Neurips 2022
♻ ☆ Spectral Maps for Learning on Subgraphs
In graph learning, maps between graphs and their subgraphs frequently arise. For instance, when coarsening or rewiring operations are present along the pipeline, one needs to keep track of the corresponding nodes between the original and modified graphs. Classically, these maps are represented as binary node-to-node correspondence matrices and used as-is to transfer node-wise features between the graphs. In this paper, we argue that simply changing this map representation can bring notable benefits to graph learning tasks. Drawing inspiration from recent progress in geometry processing, we introduce a spectral representation for maps that is easy to integrate into existing graph learning models. This spectral representation is a compact and straightforward plug-in replacement and is robust to topological changes of the graphs. Remarkably, the representation exhibits structural properties that make it interpretable, drawing an analogy with recent results on smooth manifolds. We demonstrate the benefits of incorporating spectral maps in graph learning pipelines, addressing scenarios where a node-to-node map is not well defined, or in the absence of exact isomorphism. Our approach bears practical benefits in knowledge distillation and hierarchical learning, where we show comparable or improved performance at a fraction of the computational cost.
♻ ☆ Rateless Autoencoder Codes: Trading off Decoding Delay and Reliability
Most of today's communication systems are designed to target reliable message recovery after receiving the entire encoded message (codeword). However, in many practical scenarios, the transmission process may be interrupted before receiving the complete codeword. This paper proposes a novel rateless autoencoder (AE)-based code design suitable for decoding the transmitted message before the noisy codeword is fully received. Using particular dropout strategies applied during the training process, rateless AE codes allow to trade off between decoding delay and reliability, providing a graceful improvement of the latter with each additionally received codeword symbol. The proposed rateless AEs significantly outperform the conventional AE designs for scenarios where it is desirable to trade off reliability for lower decoding delay.
comment: 6 pages, 7 figures, to appear at IEEE ICC 2023
♻ ☆ Difformer: Empowering Diffusion Models on the Embedding Space for Text Generation
Diffusion models have achieved state-of-the-art synthesis quality on both visual and audio tasks, and recent works further adapt them to textual data by diffusing on the embedding space. In this paper, we conduct systematic studies and analyze the challenges between the continuous data space and the embedding space which have not been carefully explored. Firstly, the data distribution is learnable for embeddings, which may lead to the collapse of the loss function. Secondly, as the norm of embeddings varies between popular and rare words, adding the same noise scale will lead to sub-optimal results. In addition, we find the normal level of noise causes insufficient training of the model. To address the above challenges, we propose Difformer, an embedding diffusion model based on Transformer, which consists of three essential modules including an additional anchor loss function, a layer normalization module for embeddings, and a noise factor to the Gaussian noise. Experiments on two seminal text generation tasks including machine translation and text summarization show the superiority of Difformer over compared embedding diffusion baselines.
♻ ☆ Personalized Decentralized Bilevel Optimization over Random Directed Networks
Personalization and decentralization are two major lines of studies to realize practical federated learning in the real world. The aim of this study is to establish a general and unified approach that can solve these two problems simultaneously. In this work, we first propose a bilevel problem that can adapt to various personalization scenarios by allowing an arbitrary choice of two parameters: a client-wise outer-parameter representing heterogeneity, and a shared inner-parameter representing homogeneity across client data distributions. We then present an algorithm that can solve this bilevel problem in a decentralized manner by estimating gradients of clients' outer-costs with respect to their outer-parameters. We show that the proposed algorithm can be extended to handle a random directed network, which is one of the most robust decentralized communication classes. The proposed method achieves state-of-the-art performance on a personalization benchmark across various communication settings.
comment: Under review
♻ ☆ Day-Ahead PV Power Forecasting Based on MSTL-TFT
In recent years, renewable energy resources have accounted for an increasing share of electricity energy.Among them, photovoltaic (PV) power generation has received broad attention due to its economic and environmental benefits.Accurate PV generation forecasts can reduce power dispatch from the grid, thus increasing the supplier's profit in the day-ahead electricity market.The power system of a PV site is affected by solar radiation, PV plant properties and meteorological factors, resulting in uncertainty in its power output.This study used multiple seasonal-trend decomposition using LOESS (MSTL) and temporal fusion transformer (TFT) to perform day-ahead PV prediction on the desert knowledge Australia solar centre (DKASC) dataset.We compare the decomposition algorithms (VMD, EEMD and VMD-EEMD) and prediction models (BP, LSTM and XGBoost, etc.) which are commonly used in PV prediction presently.The results show that the MSTL-TFT method is more accurate than the aforementioned methods, which have noticeable improvement compared to other recent day-ahead PV predictions on desert knowledge Australia solar centre (DKASC).
♻ ☆ Inference on the Optimal Assortment in the Multinomial Logit Model
Assortment optimization has received active explorations in the past few decades due to its practical importance. Despite the extensive literature dealing with optimization algorithms and latent score estimation, uncertainty quantification for the optimal assortment still needs to be explored and is of great practical significance. Instead of estimating and recovering the complete optimal offer set, decision-makers may only be interested in testing whether a given property holds true for the optimal assortment, such as whether they should include several products of interest in the optimal set, or how many categories of products the optimal set should include. This paper proposes a novel inferential framework for testing such properties. We consider the widely adopted multinomial logit (MNL) model, where we assume that each customer will purchase an item within the offered products with a probability proportional to the underlying preference score associated with the product. We reduce inferring a general optimal assortment property to quantifying the uncertainty associated with the sign change point detection of the marginal revenue gaps. We show the asymptotic normality of the marginal revenue gap estimator, and construct a maximum statistic via the gap estimators to detect the sign change point. By approximating the distribution of the maximum statistic with multiplier bootstrap techniques, we propose a valid testing procedure. We also conduct numerical experiments to assess the performance of our method.
♻ ☆ FedPass: Privacy-Preserving Vertical Federated Deep Learning with Adaptive Obfuscation
Vertical federated learning (VFL) allows an active party with labeled feature to leverage auxiliary features from the passive parties to improve model performance. Concerns about the private feature and label leakage in both the training and inference phases of VFL have drawn wide research attention. In this paper, we propose a general privacy-preserving vertical federated deep learning framework called FedPass, which leverages adaptive obfuscation to protect the feature and label simultaneously. Strong privacy-preserving capabilities about private features and labels are theoretically proved (in Theorems 1 and 2). Extensive experimental result s with different datasets and network architectures also justify the superiority of FedPass against existing methods in light of its near-optimal trade-off between privacy and model performance.
comment: 6 figures, 9 tables
♻ ☆ Learning to reject meets OOD detection: Are all abstentions created equal?
Learning to reject (L2R) and out-of-distribution (OOD) detection are two classical problems, each of which involve detecting certain abnormal samples: in L2R, the goal is to detect "hard" samples on which to abstain, while in OOD detection, the goal is to detect "outlier" samples not drawn from the training distribution. Intriguingly, despite being developed in parallel literatures, both problems share a simple baseline: the maximum softmax probability (MSP) score. However, there is limited understanding of precisely how these problems relate. In this paper, we formally relate these problems, and show how they may be jointly solved. We first show that while MSP is theoretically optimal for L2R, it can be theoretically sub-optimal for OOD detection in some important practical settings. We then characterize the Bayes-optimal classifier for a unified formulation that generalizes both L2R and OOD detection. Based on this, we design a plug-in approach for learning to abstain on both inlier and OOD samples, while constraining the total abstention budget. Experiments on benchmark OOD datasets demonstrate that our approach yields competitive classification and OOD detection performance compared to baselines from both literatures.
♻ ☆ Personalized Subgraph Federated Learning
Subgraphs of a larger global graph may be distributed across multiple devices, and only locally accessible due to privacy restrictions, although there may be links between subgraphs. Recently proposed subgraph Federated Learning (FL) methods deal with those missing links across local subgraphs while distributively training Graph Neural Networks (GNNs) on them. However, they have overlooked the inevitable heterogeneity between subgraphs comprising different communities of a global graph, consequently collapsing the incompatible knowledge from local GNN models. To this end, we introduce a new subgraph FL problem, personalized subgraph FL, which focuses on the joint improvement of the interrelated local GNNs rather than learning a single global model, and propose a novel framework, FEDerated Personalized sUBgraph learning (FED-PUB), to tackle it. Since the server cannot access the subgraph in each client, FED-PUB utilizes functional embeddings of the local GNNs using random graphs as inputs to compute similarities between them, and use the similarities to perform weighted averaging for server-side aggregation. Further, it learns a personalized sparse mask at each client to select and update only the subgraph-relevant subset of the aggregated parameters. We validate our FED-PUB for its subgraph FL performance on six datasets, considering both non-overlapping and overlapping subgraphs, on which it significantly outperforms relevant baselines.
♻ ☆ GeneFormer: Learned Gene Compression using Transformer-based Context Modeling
With the development of gene sequencing technology, an explosive growth of gene data has been witnessed. And the storage of gene data has become an important issue. Traditional gene data compression methods rely on general software like G-zip, which fails to utilize the interrelation of nucleotide sequence. Recently, many researchers begin to investigate deep learning based gene data compression method. In this paper, we propose a transformer-based gene compression method named GeneFormer. Specifically, we first introduce a modified transformer structure to fully explore the nucleotide sequence dependency. Then, we propose fixed-length parallel grouping to accelerate the decoding speed of our autoregressive model. Experimental results on real-world datasets show that our method saves 29.7% bit rate compared with the state-of-the-art method, and the decoding speed is significantly faster than all existing learning-based gene compression methods.
♻ ☆ FE-TCM: Filter-Enhanced Transformer Click Model for Web Search
Constructing click models and extracting implicit relevance feedback information from the interaction between users and search engines are very important to improve the ranking of search results. Using neural network to model users' click behaviors has become one of the effective methods to construct click models. In this paper, We use Transformer as the backbone network of feature extraction, add filter layer innovatively, and propose a new Filter-Enhanced Transformer Click Model (FE-TCM) for web search. Firstly, in order to reduce the influence of noise on user behavior data, we use the learnable filters to filter log noise. Secondly, following the examination hypothesis, we model the attraction estimator and examination predictor respectively to output the attractiveness scores and examination probabilities. A novel transformer model is used to learn the deeper representation among different features. Finally, we apply the combination functions to integrate attractiveness scores and examination probabilities into the click prediction. From our experiments on two real-world session datasets, it is proved that FE-TCM outperforms the existing click models for the click prediction.
♻ ☆ Learning to Reject with a Fixed Predictor: Application to Decontextualization
We study the problem of classification with a reject option for a fixed predictor, applicable in natural language processing. We introduce a new problem formulation for this scenario, and an algorithm minimizing a new surrogate loss function. We provide a complete theoretical analysis of the surrogate loss function with a strong $H$-consistency guarantee. For evaluation, we choose the decontextualization task, and provide a manually-labelled dataset of $2\mathord,000$ examples. Our algorithm significantly outperforms the baselines considered, with a $\sim\!\!25\%$ improvement in coverage when halving the error rate, which is only $\sim\!\! 3 \%$ away from the theoretical limit.
♻ ☆ Student-centric Model of Learning Management System Activity and Academic Performance: from Correlation to Causation
In recent years, there is a lot of interest in modeling students' digital traces in Learning Management System (LMS) to understand students' learning behavior patterns including aspects of meta-cognition and self-regulation, with the ultimate goal to turn those insights into actionable information to support students to improve their learning outcomes. In achieving this goal, however, there are two main issues that need to be addressed given the existing literature. Firstly, most of the current work is course-centered (i.e. models are built from data for a specific course) rather than student-centered; secondly, a vast majority of the models are correlational rather than causal. Those issues make it challenging to identify the most promising actionable factors for intervention at the student level where most of the campus-wide academic support is designed for. In this paper, we explored a student-centric analytical framework for LMS activity data that can provide not only correlational but causal insights mined from observational data. We demonstrated this approach using a dataset of 1651 computing major students at a public university in the US during one semester in the Fall of 2019. This dataset includes students' fine-grained LMS interaction logs and administrative data, e.g. demographics and academic performance. In addition, we expand the repository of LMS behavior indicators to include those that can characterize the time-of-the-day of login (e.g. chronotype). Our analysis showed that student login volume, compared with other login behavior indicators, is both strongly correlated and causally linked to student academic performance, especially among students with low academic performance. We envision that those insights will provide convincing evidence for college student support groups to launch student-centered and targeted interventions that are effective and scalable.
comment: Revisions recommended and submission to new venue
♻ ☆ Variational Causal Inference
Estimating an individual's potential outcomes under counterfactual treatments is a challenging task for traditional causal inference and supervised learning approaches when the outcome is high-dimensional (e.g. gene expressions, impulse responses, human faces) and covariates are relatively limited. In this case, to construct one's outcome under a counterfactual treatment, it is crucial to leverage individual information contained in its observed factual outcome on top of the covariates. We propose a deep variational Bayesian framework that rigorously integrates two main sources of information for outcome construction under a counterfactual treatment: one source is the individual features embedded in the high-dimensional factual outcome; the other source is the response distribution of similar subjects (subjects with the same covariates) that factually received this treatment of interest.
♻ ☆ WISE: Wavelet Transformation for Boosting Transformers' Long Sequence Learning Ability
Transformer and its variants are fundamental neural architectures in deep learning. Recent works show that learning attention in the Fourier space can improve the long sequence learning capability of Transformers. We argue that wavelet transform shall be a better choice because it captures both position and frequency information with a linear time complexity. Therefore, in this paper, we systematically study the synergy between wavelet transform and Transformers. Specifically, we focus on a new paradigm WISE, which replaces the attention in Transformers by (1) applying forward wavelet transform to project the input sequences to multi-resolution bases, (2) conducting non-linear transformations in the wavelet coefficient space, and (3) reconstructing the representation in input space via backward wavelet transform. Extensive experiments on the Long Range Arena benchmark demonstrate that learning attention in the wavelet space using either fixed or adaptive wavelets can consistently improve Transformer's performance and also significantly outperform Fourier-based methods.
Multimedia 6
☆ A Closer Look into Recent Video-based Learning Research: A Comprehensive Review of Video Characteristics, Tools, Technologies, and Learning Effectiveness
People increasingly use videos on the Web as a source for learning. To support this way of learning, researchers and developers are continuously developing tools, proposing guidelines, analyzing data, and conducting experiments. However, it is still not clear what characteristics a video should have to be an effective learning medium. In this paper, we present a comprehensive review of 257 articles on video-based learning for the period from 2016 to 2021. One of the aims of the review is to identify the video characteristics that have been explored by previous work. Based on our analysis, we suggest a taxonomy which organizes the video characteristics and contextual aspects into eight categories: (1) audio features, (2) visual features, (3) textual features, (4) instructor behavior, (5) learners activities, (6) interactive features (quizzes, etc.), (7) production style, and (8) instructional design. Also, we identify four representative research directions: (1) proposals of tools to support video-based learning, (2) studies with controlled experiments, (3) data analysis studies, and (4) proposals of design guidelines for learning videos. We find that the most explored characteristics are textual features followed by visual features, learner activities, and interactive features. Text of transcripts, video frames, and images (figures and illustrations) are most frequently used by tools that support learning through videos. The learner activity is heavily explored through log files in data analysis studies, and interactive features have been frequently scrutinized in controlled experiments. We complement our review by contrasting research findings that investigate the impact of video characteristics on the learning effectiveness, report on tasks and technologies used to develop tools that support learning, and summarize trends of design guidelines to produce learning videos
☆ Zero3D: Semantic-Driven Multi-Category 3D Shape Generation
Semantic-driven 3D shape generation aims to generate 3D objects conditioned on text. Previous works face problems with single-category generation, low-frequency 3D details, and requiring a large number of paired datasets for training. To tackle these challenges, we propose a multi-category conditional diffusion model. Specifically, 1) to alleviate the problem of lack of large-scale paired data, we bridge the text, 2D image and 3D shape based on the pre-trained CLIP model, and 2) to obtain the multi-category 3D shape feature, we apply the conditional flow model to generate 3D shape vector conditioned on CLIP embedding. 3) to generate multi-category 3D shape, we employ the hidden-layer diffusion model conditioned on the multi-category shape vector, which greatly reduces the training time and memory consumption.
☆ Sport Task: Fine Grained Action Detection and Classification of Table Tennis Strokes from Videos for MediaEval 2022
Sports video analysis is a widespread research topic. Its applications are very diverse, like events detection during a match, video summary, or fine-grained movement analysis of athletes. As part of the MediaEval 2022 benchmarking initiative, this task aims at detecting and classifying subtle movements from sport videos. We focus on recordings of table tennis matches. Conducted since 2019, this task provides a classification challenge from untrimmed videos recorded under natural conditions with known temporal boundaries for each stroke. Since 2021, the task also provides a stroke detection challenge from unannotated, untrimmed videos. This year, the training, validation, and test sets are enhanced to ensure that all strokes are represented in each dataset. The dataset is now similar to the one used in [1, 2]. This research is intended to build tools for coaches and athletes who want to further evaluate their sport performances.
comment: MediaEval 2022 Workshop, Jan 2023, Bergen, Norway. arXiv admin note: substantial text overlap with arXiv:2112.11384
☆ Towards Better Quality of Experience in\\HTTP Adaptive Streaming
HTTP Adaptive Streaming (HAS) is nowadays a popular solution for multimedia delivery. The novelty of HAS lies in the possibility of continuously adapting the streaming session to current network conditions, facilitated by Adaptive Bitrate (ABR) algorithms. Various popular streaming and Video on Demand services such as Netflix, Amazon Prime Video, and Twitch use this method. Given this broad consumer base, ABR algorithms continuously improve to increase user satisfaction. The insights for these improvements are, among others, gathered within the research area of Quality of Experience (QoE). Within this field, various researchers have dedicated their works to identifying potential impairments and testing their impact on viewers' QoE. Two frequently discussed visual impairments influencing QoE are stalling events and quality switches. So far, it is commonly assumed that those stalling events have the worst impact on QoE. This paper challenged this belief and reviewed this assumption by comparing stalling events with multiple quality and high amplitude quality switches. Two subjective studies were conducted. During the first subjective study, participants received a monetary incentive, while the second subjective study was carried out with volunteers. The statistical analysis demonstrated that stalling events do not result in the worst degradation of QoE. These findings suggest that a reevaluation of the effect of stalling events in QoE research is needed. Therefore, these findings may be used for further research and to improve current adaptation strategies in ABR algorithms.
♻ ☆ Dynamic Storyboard Generation in an Engine-based Virtual Environment for Video Production
Amateurs working on mini-films and short-form videos usually spend lots of time and effort on the multi-round complicated process of setting and adjusting scenes, plots, and cameras to deliver satisfying video shots. We present Virtual Dynamic Storyboard (VDS) to allow users storyboarding shots in virtual environments, where the filming staff can easily test the settings of shots before the actual filming. VDS runs on a "propose-simulate-discriminate" mode: Given a formatted story script and a camera script as input, it generates several character animation and camera movement proposals following predefined story and cinematic rules to allow an off-the-shelf simulation engine to render videos. To pick up the top-quality dynamic storyboard from the candidates, we equip it with a shot ranking discriminator based on shot quality criteria learned from professional manual-created data. VDS is comprehensively validated via extensive experiments and user studies, demonstrating its efficiency, effectiveness, and great potential in assisting amateur video production.
comment: Project page: https://virtualfilmstudio.github.io/
♻ ☆ Enhancing Dyadic Relations with Homogeneous Graphs for Multimodal Recommendation
User interaction data in recommender systems is a form of dyadic relation that reflects the preferences of users with items. Learning the representations of these two discrete sets of objects, users and items, is critical for recommendation. Recent multimodal recommendation models leveraging multimodal features (e.g., images and text descriptions) have been demonstrated to be effective in improving recommendation accuracy. However, state-of-the-art models enhance the dyadic relations between users and items by considering either user-user or item-item relations, leaving the high-order relations of the other side (i.e., users or items) unexplored. Furthermore, we experimentally reveal that the current multimodality fusion methods in the state-of-the-art models may degrade their recommendation performance. That is, without tainting the model architectures, these models can achieve even better recommendation accuracy with uni-modal information. On top of the finding, we propose a model that enhances the dyadic relations by learning Dual RepresentAtions of both users and items via constructing homogeneous Graphs for multimOdal recommeNdation. We name our model as DRAGON. Specifically, DRAGON constructs the user-user graph based on the commonly interacted items and the item-item graph from item multimodal features. It then utilizes graph learning on both the user-item heterogeneous graph and the homogeneous graphs (user-user and item-item) to obtain the dual representations of users and items. To capture information from each modality, DRAGON employs a simple yet effective fusion method, attentive concatenation, to derive the representations of users and items. Extensive experiments on three public datasets and seven baselines show that DRAGON can outperform the strongest baseline by 22.03% on average. Various ablation studies are conducted on DRAGON to validate its effectiveness.
comment: modify the format